2021 Article 3612

Journal of Ambient Intelligence and Humanized Computing (2023) 14:8459–8486
https://doi.org/10.1007/s12652-021-03612-z
ORIGINAL RESEARCH
Artificial intelligence in disease diagnosis: a systematic literature

review, synthesizing framework and future research agenda
Yogesh Kumar1 · Apeksha Koul2 · Ruchi Singla3 · Muhammad Fazal Ijaz4
Received: 4 December 2020 / Accepted: 18 November 2021 / Published online: 13 January 2022
© The Author(s), under exclusive licence to Springer-Verlag GmbH Germany, part of Springer Nature 2021
Abstract
Artificial intelligence can assist providers in a variety of patient care and intelligent health systems. Artificial intelligence
techniques ranging from machine learning to deep learning are prevalent in healthcare for disease diagnosis, drug discovery,
and patient risk identification. Numerous medical data sources are required to perfectly diagnose diseases using artificial
intelligence techniques, such as ultrasound, magnetic resonance imaging, mammography, genomics, computed tomography
scan, etc. Furthermore, artificial intelligence primarily enhanced the infirmary experience and sped up preparing patients
to continue their rehabilitation at home. This article covers the comprehensive survey based on artificial intelligence tech-
niques to diagnose numerous diseases such as Alzheimer, cancer, diabetes, chronic heart disease, tuberculosis, stroke and
cerebrovascular, hypertension, skin, and liver disease. We conducted an extensive survey including the used medical imag-
ing dataset and their feature extraction and classification process for predictions. Preferred reporting items for systematic
reviews and Meta-Analysis guidelines are used to select the articles published up to October 2020 on the Web of Science,
Scopus, Google Scholar, PubMed, Excerpta Medical Database, and Psychology Information for early prediction of distinct
kinds of diseases using artificial intelligence-based techniques. Based on the study of different articles on disease diagnosis,
the results are also compared using various quality parameters such as prediction rate, accuracy, sensitivity, specificity, the
area under curve precision, recall, and F1-score.
Keywords Artificial intelligence · Alzheimer · Cancer disease · Chronic disease · Heart disease · Tuberculosis
1 Introduction
Healthcare is shaping up in front of our eyes with advances

in digital healthcare technologies such as artificial intel-
* Muhammad Fazal Ijaz ligence (AI), 3D printing, robotics, nanotechnology, etc.
[email protected] Digitized healthcare presents numerous opportunities for
Yogesh Kumar reducing human errors, improving clinical outcomes, track-
[email protected] ing data over time, etc. AI methods from machine learn-
Apeksha Koul ing to deep learning assume a crucial function in numerous
[email protected] well-being-related domains, including improving new clini-
Ruchi Singla cal systems, patient information and records, and treating
[email protected] various illnesses (Usyal et al. 2020; Zebene et al. 2019).
The AI techniques are also most efficient in identifying
1
Department of Computer Engineering, Indus Institute the diagnosis of different types of diseases. The presence
of Technology and Engineering, Indus University,
Ahmedabad 382115, India of computerized reasoning (AI) as a method for improved
2 medical services offers unprecedented occasions to recuper-
Shri Mata Vaishno Devi University, Jammu, India
ate patient and clinical group results, decrease costs, etc.
3
Department of Research, Innovations, Sponsored Projects The models used are not limited to computerization, such as
and Entrepreneurship, CGC Landran, Mohali, India
providing patients, “family” (Musleh et al. 2019; Dabowsa
4
Department of Intelligent Mechatronics Engineering, Sejong et al. 2017), and medical service experts for data creation
University, Seoul 05006, South Korea
13
Vol.:(0123456789)
8460 Y. Kumar et al.
and suggestions as well as disclosure of data for shared eval- the data designs. They proposed a five-phase machine learn-
uation building. AI can also help to recognize the precise ing pipeline that further arranged each stage in various sub
demographics or environmental areas where the frequency levels. They built a classifier framework alongside infor-
of illness or high-risk behaviors exists. Researchers have mation change and highlighted choice procedures inserted
effectively used deep learning classifications in diagnostic inside a test and information investigation plan. Skaane et al.
approaches to computing links between the built environ- (2013) enquired the property of digital breast tomosynthesis
ment and obesity frequency (Bhatt et al. 2019; Plawiak et al. on period and detected cancer in residents based screen-
2018). ing. They did a self-determining dual analysis examination
AI algorithms must be trained on population-represent- by engaging ladies of 50–69 years and comparing full-field
ative information to accomplish presentation levels essen- digitized mammography plus data building tool with full-
tial for adaptable “accomplishment”. Trends, such as the field digital mammography. Accumulation of the data build-
charge for putting away and directing realities, information ing tool resulted in a non-significant enhancement in sensi-
collection through electronic well-being records (Minaee tivity by 76.2% and a significant increase by 96.4%. Tigga
et al. 2020; Kumar 2020), and exponential client state of et al. (2020) aimed to assess the diabetic risk among the
information, have made a data-rich medical care biological patients based on their lifestyle, daily routines, health prob-
system. This enlargement in health care data struggles with lems, etc. They experimented on 952 collected via an offline
the lack of well-organized mechanisms for integrating and and online questionnaire. The same was applied to the Pima
reconciling these data ahead of their current silos. However, Indian Diabetes database. The random forest classifier stood
numerous frameworks and principles facilitate summation out to be the best algorithm. Alfian et al. (2018) presented a
and accomplish adequate data quantity for AI (Vasal et al. personalized healthcare monitoring system using Bluetooth-
2020). The challenges in the operational dynamism of AI based sensors and real-time data processing. It gathers the
technologies in healthcare systems are immeasurable despite user’s vital signs data such as blood pressure, heart rate,
the information that this is one of the most vital expansion weight, and blood glucose from sensor nodes to a smart-
areas in biomedical research (Kumar et al. 2020). The AI phone. Katherine et al. (2019) gave an overview of the types
commune must build an integrated best practice method of data encountered during the setting of chronic disease.
for execution and safeguarding by incorporating active best Using various machine learning algorithms, they explained
practices of principled inclusivity, software growth, imple- the extreme value theory to better quantify severity and risk
mentation science, and individual–workstation interaction. in chronic disease. Gonsalves et al. (2019) aimed to pre-
At the same time, AI applications have an enormous ability dict coronary heart disease using historical medical data via
to work on patient outcomes. Simultaneously, they could machine learning technology. The presented work supported
make significant hazards regarding inappropriate patient three supervised learning techniques named Naïve Bayes,
risk assessment, diagnostic inaccuracy, healing recommen Support vector machine, and Decision tree to find the corre-
dations, privacy breaches, and other harms (Gouda et al. lations in coronary heart disease, which would help improve
2020; Khan and Member 2020). the prediction rate. The authors worked on the South African
Researchers have used various AI-based techniques such Heart Disease dataset of 462 instances and machine learn-
as machine and deep learning models to detect the diseases ing techniques using 10-fold cross-validation. Momin et al.
such as skin, liver, heart, alzhemier, etc. that need to be (2019) proposed a secure internet of things-based healthcare
diagnosed early. Hence, in related work, the techniques like system utilizing a body sensor network called body sensor
Boltzmann machine, K nearest neighbour (kNN), support network care to accomplish the requirements efficiently. The
vector machine (SVM), decision tree, logistic regression, sensors used analogue to digital converter, Microcontroller,
fuzzy logic, and artificial neural network to diagnose the dis- cloud database, network, etc. A study by Ijaz et al. (2018)
eases are presented along with their accuracies. For example, has used IoT for a healthcare monitoring system for diabe-
a research study by Dabowsa et al. (2017) used a backpropa- tes and hypertension patients at home and used personal
gation neural network in diagnosing skin disease to achieve healthcare devices that perceive and estimate a persons’
the highest level of accuracy. The authors used real-world biomedical signals. The system can notify health personnel
data collected from the dermatology department. Ansari in real-time when patients experience emergencies. Shabut
et al. (2011) used a recurrent neural network (RNN) to diag- et al. (2018) introduced an examination to improve a smart,
nose liver disease hepatitis virus and achieved 97.59%, while versatile, empowered master to play out a programmed
a feed-forward neural network achieved 100%. Owasis et al. discovery of tuberculosis. They applied administered AI
(2019) got 97.057 area under the curve by using residual method to achieve parallel grouping from eighteenth lower
neural network and long short-term memory to diagnose request shading minutes. Their test indicated a precision
gastrointestinal disease. Khan and Member (2020) intro- of 98.4%, particularly for the tuberculosis antigen explicit
duced a computerized arrangement framework to recover counteracting agent identification on the portable stage.
13
Artificial intelligence in disease diagnosis: a systematic literature review, synthesizing… 8461
Tran et al. (2019) provided the global trends and develop- Forest, K-Mean clustering, RNN, Convolutional neural
ments of artificial intelligence applications related to stroke networks (CNN), Deep-CNN, Generative Adversarial
and heart diseases to identify the research gaps and suggest Networks (GAN), and Long short-term memory (LSTM)
future research directions. Matusoka et al. (2020) stated that and many others for various disease detection system
the mindfulness, treatment, and control of hypertension are (Owasis et al. 2019; Nithya et al. 2020). We conducted an
the most significant in overcoming stroke and cardiovascu- extensive survey based on the machine and deep learning
lar infection. Rathod et al. (2018) proposed an automated models for disease diagnosis. The study covers the review
image-based retrieval system for skin disease using machine of various diseases and their diagnostic methods using AI
learning classification. Srinivasu et al. (2021a, b) proposed techniques. This contribution explains by addressing the
an effective model that can help doctors diagnose skin dis- four research questions: RQ1. What is the state-of-the-art
ease efficiently. The system combined neural networks with research for AI in disease diagnosis? RQ2. What are the
MobileNet V2 and Long Short Term Memory (LSTM) with various types of diseases wherein AI is applied? RQ3.
an accuracy rate of 85%, exceeding other state-of-the-art What are the emergent limitations and challenges that the
deep models of deep learning neural networks. This sys- literature advances for this research area? RQ4.What are
tem utilized the technique to analyse, process, and relegate the future avenues in healthcare that might benefit from
the image data predicted based on various features. As a the application of AI? The rest of the work is organized
result, it gave more accuracy and generated faster results as into various sections. Initially, a brief description of AI in
compared to the traditional methods. Uehara et al. (2018) healthcare and disease diagnosis using multiple machines
worked at the Japanese extremely chubby patients utilizing and deep learning techniques is given in Sect. 1. Then, it
artificial brainpower with rule extraction procedure. They is named an introduction that includes Fig. 1 to describe
had 79 Non-alcoholic steatohepatitis, and 23 non- Non- all the papers taken from different organized sources for
alcoholic steatohepatitis patients analyse d to make the various diseases in the contribution sub-section. Materi-
desired model. They accomplished the prescient exactness als and Methods is named as Sect. 2, which includes the
by 79.2%. Ijaz et al. (2020) propose a cervical cancer pre- quality assessment and the investigation part regarding AI
diction model for early prediction of cervical cancer using techniques and applications. Section 3 covers symptoms
risk factors as inputs. The authors utilize several machine of diseases and challenges to diagnostics, a framework for
learning approaches and outlier detection for different pre- AI in disease detection modelling, and various AI applica-
processing tasks. Srinivasu et al. (2021a, b) used an AW- tions in healthcare. Section 4 includes the reported work of
HARIS algorithm to perform automated segmentation of CT multiple diseases and the comparative analysis of differ-
scan images to identify abnormalities in the human liver. It ent techniques with the used dataset, applied machine and
is observed that the proposed approach has outperformed in deep learning methods with computed outcomes in terms
the majority of the cases with an accuracy of 78%. of various parameters such as accuracy, sensitivity, speci-
To fully understand how AI assists in the diagnosis and ficity, the area under the curve, and F-score. In Sect. 5, the
prediction of a disease, it is essential to understand the discussion part is covered that answers the investigation
use and applicability of diverse techniques such as SVM, part mentioned in Sect. 2. Finally, in Sect. 6, the work that
KNN, Naïve Bayes, Decision Tree, Ada Boost, Random helps researchers chooses the best approach for diagnos-
ing the diseases is concluded along with the future scope.
Fig. 1 Distribution of published

papers for diseases diagnosis
using artificial intelligence
techniques
13
1.1 Contribution n-space, and qi and pi are the Euclidean vectors starting

from the origin of the space (Zaar et al. 2020).
Diseases usually are quantified by signs and symptoms. A √
sign is identified as an objective appearance of a disease d(p, q) = d(q, p) = (q1 − p1 )2 + (q2 + p2 )2 + ⋯ + (qn + pn )2
√
that doctors can specify, whereas a symptom is a particular √ n
√∑ (3)
indication of the patient’s illness (Plawiak et al. 2018). Thus, = √ (qi − pi )2
i=1
every disease has various signs and symptoms, such as fever,
which is found in countless conditions. Whereas regression is used to determine the relation-
As shown in Fig. 1, the number of papers reviewed under ship between independent and dependent variables. The
preferred reporting items for systematic reviews and Meta- equation Y represents it is equal to an X plus b, where Y
Analysis (PRISMA) guidelines for different types of dis- is the dependent variable, an is the slope of the regression
eases using AI from the year 2009 to the year 2020. The pre- equation, x is the independent variable, and b is constant
sent work emphasizes various diseases and their diagnostics (Kolkur et al. 2018)
measures using machine and deep learning classifications.
To the best of our knowledge, most of the past work focused Y = a + bX + 𝜖, (4)
on disease diagnostics for one or two disease prediction sys- where Y is the dependent variable, X is the independent
tems. Hence, the present study explores ten different disease variable; a is the intercept, b is the slope and is the residual
symptoms and their detection using AI techniques. Further- error, Naïve Bayes which provides a way of calculating the
more, this paper is unique, as it contains an elaborate discus- posterior probability, P (c | x) from P(c), P(x) and P(x | c).
sion about various disease diagnoses and predictions based Naïve Bayes classifier assumes that the effect of the value
upon the extensive survey conducted for detection methods. of an attribute (x) on a given class (c) is independent of the
values of other predictors (Spann et al. 2020)
2 Materials and methods P(x|c)P(c)

P(c|x) = , (5)
P(x)
We have directed this review according to the preferred
reporting items for systematic reviews and Meta-Analysis P(c | x) is the posterior probability of class given attrib-
guidelines. The survey offers the readers wide-ranging ute, P(x | c) is the likelihood which is the probability of the
knowledge of the literature on AI (decision tree, which attribute given class, P(x) is the prior probability of attribute,
breaks down the dataset into smaller subsets and to build P(c) is the prior probability of a class, k-means ( Fujita et al.
it, two types of entropy using frequencies are calculated in 2020) which is used to define k centers, one for each cluster
which X, S is a discrete random variable which occurs with and these centres should be placed far away from each other.
probability p(i),…. p(c) and logarithm with base 2 gives the This algorithm also aims at minimizing an objective func-
unit of bits or Shannons where entropy using the frequency tion which is known as squared error function, given by :
table of one attribute is given as (Sabottke and Spieler 2020) c
∑ ci (
∑ )
|| || 2
c
∑
J(v) = ||xi − vj || ;
|| || (6)
i=1 j=1
E(S) = −pi log2 pi , (1)
i=1
||xi-vj|| is the Euclidean distance between x i-vj, Ci is the num-
and entropy using the frequency table of two attributes is ber of data points in ith cluster, C is the number of cluster
given as center’s, convolution neural network which is a type of feed-
∑ forward artificial neural network in which the connectivity
E(S, X) = P(c)E(c) (2) pattern between its neurons is inspired by the organization of
c∈X the animal visual cortex. Convolution is the first step in the
K-nearest neighbour algorithm is a supervised machine process that convolution neural network undergoes (Zhang
learning technique that is used to solve classification issues et al. 2019)
as well as to calculate the distance between the test data ∞
∫
and the input to give the prediction by using Euclidean dis- def
(f ∗ g)(t) = f (𝜏) ⋅ g(x − 𝜏)d𝜏 (7)
tance formula in which p, q are the two points in Euclidean
−∞
13
where (f*g)(t) = functions that are being convoluted, t = real to represent the techniques used, and research design to
number variable of functions f and g, g(𝜏 ) = convolution of analyse the results that are obtained. These characteristics
time function, 𝜏 ′ = first derivative of tau function, a recurrent assisted us in carrying out the research study very quickly,
neural network which is used for handling sequential data without wasting time on irrelevant or unnecessary searches
and its formula in which h(t) is a function f of the previously and investigations. The standards for inclusion and exclu-
hidden state h(t − 1) and the current input x(t). The theta are sion are developed by the requirements of the problem of
the parameters of the function f is (Yang et al. 2020) an article.
h(t) = f (h(t−1) , xt ;𝜃) (8) 2.1 Quality assessment

Boltzmann machine, which optimizes the weights, a
quantity related to the particular problem. Its main objec- Research articles included in this review are identified using
tive is to maximize the Consensus function (CF), which is several quality evaluation constraints. The significance of the
given by the following formula (Zhou et al. 2019) study is assessed based on inclusion and exclusion stand-
ards. All research articles included for review encompass
∑∑
F= wij ui uj machine or deep learning-based prediction models for auto-
(9)
i j≤i matically detecting and diagnosing diseases. Each research
work incorporated in this study carried empirical research
where Ui and Uj are the set of units, wij is the fixed weight, and had experimental outcomes. The description of these
gradient descent which is an iterative process and is formu- research articles is stated in a separate subsection entitled
lated by (Chang et al. 2018) literature survey.
𝜃 1 = 𝜃 0 − 𝛼∇J(𝜃) evaluated at 𝜃 0 (10) The comprehensive selection of research papers is car-
ried out in four phases: (1) identifying where records are
where 𝜃1 is the next position, 𝜃0 is the current position, 𝛼 is identified through various databases. At this phase, we
the small step, ∇J(𝜃) is the direction of fastest increase) in must do the searches we’ve planned through the abstract
healthcare (Zhang et al. 2017). The extensive survey also and citation databases we’ve chosen. Take note of how
promotes expounding prevailing knowledge gaps and sub- many results the searches returned. We can also include
sequent identification of paths for future research (Lin et al. data found in other places, such as Google Scholar or the
2019). The current study reformed the structure, which pro- reference lists of related papers. Then, in one citation man-
duced wide-ranging article valuation standards from earlier agement application, aggregate all of the records retrieved
published articles. Articles incorporated in our research from the searches. Keep in mind that each database has
are selected using keywords like “Artificial Intelligence”, its own set of rules for searching for terms of interest and
“Disease Detection”, “Disease diagnosis using machine combining keywords for a more efficient search. As a
learning”, “Disease diagnosis using deep learning”, “Artifi- result, our search technique may vary significantly depend-
cial Intelligence in Healthcare”, and combinations of these ing on the database, (2) screening the selection process is
keywords. In addition, research articles associated with the done transparently by reporting on decisions made at vari-
applications of AI-based techniques in predicting diseases ous stages of the systematic review. One of the investiga-
and diagnosing them are incorporated for review. Table 1 tors reviews the title and abstract of each record to see if
lists the publications that are included or omitted based on the publication provides information that might be useful
a variety of criteria such as time, studies to define how old or relevant to the systematic review. In certain situations,
papers/articles can be accessed, the problem on which the the title and abstract screening is done by two investiga-
article is based, comparative analysis of the work, methods tors. They don’t split the job amongst themselves! Each
Table 1 Inclusion and exclusion parameters

S. no. Parameters Inclusion standards Exclusion standards
1. Period Research works conducted between 2009 and 2020 Articles published before 2009
2. Investigations Research works focusing on disease diagnosis using AI Research works focusing other than disease diagnosis
3. Comparator Research studies aiming to detect the disease Research works making predictive models other than
detecting diseases
4. Methodology Research articles using ML/DL methods Research articles using methods other than ML/DL
5. Design of Study Original articles comprising of experimental results Review articles, case studies, Patents
Language other than English
13
investigator screens every title and abstract, and then their information for articles on AI in health informatics in pre-
judgments are compared. If one of them decides to leave vious studies. As previously explained, articles are chosen
out an item that the other thinks should be included, they based on specified inclusion and exclusion criteria (Zhang
may go over the entire text together and come to a com- et al. 2017). These were derived from (Behera et al. 2019),
mon conclusion. They can also enlist the help of a third where the authors established and accepted the variations.
party (usually the project manager or main investigator) to To better understand the state of research on AI in disease
decide whether or not the study should be included. Make detection, peer-reviewed papers are cited. The current
sure that the most acceptable justification for excluding review suggests that AI and healthcare have developed a
an item is chosen. (3) Eligibility we study the complete present synergy.
contents of the articles that cleared the title and abstract
screening to see whether they may assist in answering our 2.2 Investigation
research topic. Two investigators do this full-text screen-
ing. Each one examines the entire content of each arti- Investigation 1: Why do we need AI?
cle before deciding whether or not to include it. We must Investigation 2: What is the impact of AI on medical diag-
note the number of articles we remove and the number nosis and treatment?
of articles under each cause for exclusion in the full-text Investigation 3: Why is AI important, and how is it used
screening, just as we did in the title/abstract screening. to analyse these diseases?
Hence, in this stage, full-text articles are assessed and Investigation 4: Which AI-based algorithm is used in dis-
then finally are included in qualitative analysis in (4) ease diagnosis?
included phase by utilizing the Preferred reporting items Investigation 5: What are the challenges faced by the
for systematic reviews and meta-analysis (PRISMA) flow- researchers while using AI models in several disease
chart as depicted in Fig. 2. In this stage, we’ll know how diagnoses?
many papers will be included in our systematic review Investigation 6: How are AI-based techniques helping
after removing irrelevant studies from the full-text screen. doctors in diagnosing diseases?
We assess how many of these studies may be included
in a quantitative synthesis, commonly known as “meta-
analysis,“ in the fourth and final screening stage. 3 Artificial intelligence in disease diagnosis
To address the RQ1, RQ2, RQ3, and RQ4, the cur-
rent survey examined the number of articles on different Detecting any irresistible ailment is nearly an afterward
disease diagnoses using AI techniques from various data movement and forestalling its spread requires ongoing data
sources, including Psychological Information, Excerpta and examination. Hence, acting rapidly with accurate data
Medica Database, Google Scholar, PubMed, Scopus, and tosses a significant effect on the lives of individuals around
Web of Science. The above sources are popular sources of the globe socially and financially (Minaee et al. 2020). The
best thing about applying AI in health care is to improve
from gathering and processing valuable data to program-
ming surgeon robots. This section expounds on the various
techniques and applications of artificial intelligence, disease
symptoms, diagnostics issues, and a framework for disease
detection modelling using learning models and AI in health-
care applications (Kumar and Singla 2021).
3.1 Framework for AI in disease detection

modelling
AI describes the capability of a machine to study the way a

human learns, e.g., through image identification and detect-
ing pattern in a problematic situation. AI in health care alters
how information gets composed, analysed, and developed for
patient care (Ali et al. 2019).
System planning is the fundamental abstract design of
the system. It includes the framework’s views, the course of
action of the framework, and how the framework carries on
Fig. 2 PRISMA flow chart underneath clear conditions. A solid grip of the framework
13
Fig. 3 Framework for disease

detection system
design can help the client realize the limits and boundaries 3.2 Medical imaging for diseases diagnosis
of the said framework. Figure 3 shows a pictorial portrayal of
the ailment recognition model using utilitarian machines and Clinical Imaging is seen to assign the arrangement of pro-
profound learning classification strategies. In pre-preparing, cedures that produce pictures of the inside part of the body.
real-world information requires upkeep and pre-preparing The procedure and cycles are used to take pictures of the
before being taken care of by the calculation (Jo et al. 2019). human body for clinical purposes, such as uncovering, ana-
Because of the justifiable explanation, real-world data regu- lysing, or looking at an injury, brokenness, and pathology
larly contains mistakes regarding the utilized measures yet (Bibault et al. 2020). Computed tomography (CT) scan out-
cannot practice such blunders. Accordingly, information pre- puts are great representations of helpful indicative imaging
preparing takes this crude information, cycles it, eliminates that encourages exact conclusion, mediation, and evaluation
errors, and spares it an extra examination. Information expe- of harms and dysfunctions that actual advisors address con-
riences a progression of steps during pre-handling (Chen sistently (Chen et al. 2017). Additional contemplates demon-
et al. 2019a, b): Information is purged by various strategies strate overuse of Imaging, for example, X-rays or magnetic
in information cleaning. These strategies involve gathering resonance imaging (MRI) for intense and complicated work,
information, such as filling the information spaces that are as shown in Table 2.
left clear or decreasing information, such as the disposal of
commas or other obscure characters. In information osmosis, 3.3 Symptoms of diseases and challenges
the information is joined from a combination of sources. The to diagnostics
information is then amended for any blend of mistakes, and
they are quickly taken care of. Information Alteration: Data The disease may be severe, persistent, cruel, or benign. Of
in this progression is standardized, which depends upon these terms, persistent and severe have to do with the inter-
the given calculation. Information standardization can be val of a disease, lethal and begin with the potential for caus-
executed utilizing several ways (Nasser et al. 2019). This ing death. Additionally, different manifestations that may
progression is obligatory in most information mining calcu- be irrelevant could post the warnings for more restorative
lations, as the information wants to be as perfect as possible. severe illness or situation. The followings are a couple of
Information is then mutual and developed. Information Less- diseases with their sign and indications for events:
ening: This progression in the strategy centers to diminish
the information to more helpful levels. Informational col- • Heart assault signs incorporate hurt, nervousness, crush-
lection and test information: The informational collection ing, or feeling of breadth in the focal point of the chest
is segregated into parts preparing and testing informational that endures more than a couple of moments; agony or
indexes. The preparation information is utilized to gauge the anxiety in different territories of the chest area; succinct-
actual examples of the data (Sarao et al. 2020). Equivalent to ness of breath; cold perspiration; heaving; or unsteadi-
information needed for preparing and testing, experimental ness (Aggarwal et al. 2020).
data is often replicated from a similar informational index. • Stroke signs incorporate facial listing, arm shortcoming,
After the model has been pre-handled, the jiffy step is to test the intricacy with discourse, quickly creating happiness
the accuracy of the framework. Systematic model: Analytical or equalization, unexpected absence of sensation or weak
displaying strategies are utilized to calculate the probability point, loss of vision, puzzlement, or agonizing torment
of a given occurrence function given commitment factors, (Lukwanto et al. 2015).
and it is very productive in illness expectation. It tends to • Reproductive wellbeing manages the signs that develop
imagine what the individual is experiencing in light of their the issues such as blood misfortune or spotting between
info indications and prior determinations (Keenan et al. periods; tingling, copying, disturbance at genital region;
2020; Rajalakshmi et al. 2018). agony or disquiet during intercourse; genuine or sore
13
Table 2 Medical imaging types with their respective descriptions

Medical imaging types Description
Radiographic imaging (Zhang et al. 2019) Radiographic imaging is utilized in the ionizing of electromagnetic radiation, for
example, X-beams to see objects
Fluoroscopy (Santroo et al. 2019) It creates ongoing pictures of the body’s interior structures that consistently contrib-
ute X-beams at a lower portion rate to give moving projection radiographs of lower
quality
Angiography (Katharine et al. 2019) Angiography is utilized to discover aneurysms, releases, blockages, new vessel
development, and arrangement of catheters and stents
DEXA (Yang et al. 2020) It is likewise called Dual X-beam Absorptiometry or bone densitometry which is
utilized for osteoporosis tests
Computed tomography (CT) (Kasasbeh et al. 2019) Computed tomography examination utilizes an immense measure of ionizing radia-
tion related to a PC to make pictures of delicate and hard tissues
Magnetic resonance imaging (Zhou et al. 2019) Magnetic resonance imaging (MRI) filtering is a clinical examination that utilizes an
excellent magnet and radiofrequency waves to create a body picture
Ultrasound imaging (Sloun et al. 2019) It utilizes high recurrence broadband sound waves in the megahertz range that are
reflected by tissue to differing degrees to deliver 3D pictures
Bone scan (Gupta et al. 2019) It is an imaging procedure that utilizes a radioactive compound to distinguish the
regions of mending within the bone
Electron microscopy (Tegunov et al. 2019) Electron microscopy is a magnifying instrument that can amplify tiny subtleties with
high settling power
Nuclear medicine (Nensa et al. 2019) Nuclear medication on an entire incorporates both the finding and treatment of
infections utilizing atomic properties
Magnetic resonance angiography scans (Fujita et al. 2020) Magnetic resonance angiography represents an attractive reverberation angiogram
that gives exceptionally itemized pictures of the veins in the body
feminine dying; extreme pelvic/stomach torment; strange • Emotional issues include nervousness, sadness, weari-
vaginal release; the sentiment of totality in the lower ness, feeling tense, flashbacks and bad dreams, lack of
mid-region; and customary pee or urinary weight (Kather engagement in daily exercises, self-destructive musings,
et al. 2019). mind flights, and fancies (Krittanawong et al. 2018).
• Breast issue side effects include areola release, abnor- • Headache issues indications (excluding ordinary strain
mal bosom delicacy or torment, bosom or areola skin cerebral pains) incorporate migraines that please unex-
changes, knot or thickening in or close to bosom or in pectedly, “the most noticeably awful migraine of your
the underarm zone (Memon et al. 2019). life”, and cerebral pain connected with extreme energy,
• Lung issue side effects include hacking of blood, suc- queasiness, heaving, and powerlessness to walk (Muel-
cinctness of breath, difficult breathing, consistent hack, ler 2020).
rehashed episodes of bronchitis or pneumonia, and puff-
ing (Ma et al. 2020). Above, we have described the variety of illness signals
• Stomach or stomach-related issue manifestations incor- and their symptoms. In contrast, illness recognition errors
porate rectal dying, blood in the stool or dark stools, in medication are reasonably regular, can have a stringent
changes in gut properties or not having the option to penalty, and are only now the foundation to materialize
control guts, stoppage, loose bowels, indigestion or heart- outstandingly in patient safety. Here we have critical issues
burn, or spewing blood (Kather et al. 2019). for various diagnostic types while detecting the particular
• Bladder issue manifestations include confounded or diseases (Chuang 2011; Park et al. 2020).
excruciating pee, incessant pee, loss of bladder control,
blood in pee, waking routinely to pee around evening • Analysis that is accidentally deferred wrong, or on the
time to pee or wetting the bed around evening time, or other hand, missed as decided from a definitive delight
spilling pee (Shkolyar et al. 2019). of more amazing data.
• Skin issue indications remember changes for skin moles, • Any fault or malfunction in the analytical course which
repetitive flushing and redness of face and neck, jaundice, is essential to a missed finding or a conceded conclu-
skin sores that do not disappear or re-establish to wellbe- sion comprises a breakdown in occasional admittance
ing, new development or moles on the skin, and thick, red to mind; elicitation or comprehension of side effects,
skin with bright patches (Rodrigues et al. 2020). images, research facility result; detailing and weighing
13
of difference investigation; and ideal development and 4.1 Diagnosis of Alzheimer’s disease

strength arrangement or appraisal.
Alzheimer’s is a disease that worsens the dementia symp-
3.4 Healthcare applications toms over several years (Zebene et al. 2019). During its early
stage, it affects memory loss, but in the end, it loses the
The healthcare system has long been an early adopter of gen- ability to carry the conservation and respond to the envi-
erally innovative technologies. Today, artificial intelligence ronment. Usyal et al. (2020) decided on the analysis of
and its subset machine and deep learning are on their way dementia in Alzheimer’s through investigating neuron pic-
to becoming a mean element in the healthcare system, from tures. They utilized the alzheimer’s disease neuroimaging
creating new health check actions to treat patient records and initiative convention that comprises T1 weighted magnetic
accounts. One of the maximum burdens physician practices resonance information for finding. The prescient shows the
today is the association and performance of organizational precision estimated the characterization models, affectabil-
tasks (Fukuda et al. 2019). By automating them, healthcare ity, and explicitness esteem. Ljubic et al. (2020) presented
institutions could help resolve the trouble and allow physi- the method to diagnose Alzheimer’s disease from electronic
cians to do their best, i.e., spend more time with patients. medical record (EMR) data. The results acquired showed
The following are the details of the artificial intelligence the accuracy by 90% on using the SCRL dataset. Soundarya
techniques in healthcare applications as shown in Table 3: et al. (2020) proposed the methodology in which descrip-
tion of shrink brain tissue is used for the ancient analysis
of Alzheimer’s disease. They have implemented various
4 Reported work machine and deep learning algorithms. The deep algorithm
has been considered the better solution provider to recognize
This section highlights the best finding for different diseases the ailment at its primary stage with reasonable accuracy.
with their diagnosis methods via machine and deep learn- Park et al. (2020) used a vast range of organizational health
ing algorithms. It covers the extensive survey on various data to test the chance of machine learning models to expect
diseases such as alzheimer’s, cancer, diabetes, chronic, heart the outlook occurrence of Alzheimer’s disease. Lin et al.
disease, tuberculosis, stroke and cerebrovascular, hyperten- (2019) proposed a method that used the spectrogram features
sion, skin and liver disease (Chui et al. 2020). extracted from speech data to identify Alzheimer’s disease.
The system used the voice data collected via the internet of
things (IoT) and transmitted to the cloud server where the
Table 3 Healthcare applications and their purpose

Healthcare applications Purpose
Analysis and disease identification (Memon et al. 2019) One of the most critical uses of the machine and profound learning calculations in
medical care is identified with the acknowledgment and investigation of sicknesses
that are estimated hard to diagnose
Drug development (Memon et al. 2019) The beginning phase of the drug identification measure is a different zone that can
greatly advance from the machine and profound learning. Solo AI is beneficial to
distinguish designs in information without giving any forecast
Customized medicine (Chatterjee et al. 2018) Medicines are most solid when they are imparted to only wellbeing factors. As of now,
doctors can lean toward a lack of conclusion or inexact danger to their patients based
on their characteristic history and the open acquired data
Digital health records (Luo et al. 2019) They are keeping up just as vital well-being records are a long and expensive cycle.
As a result, they have assumed an important function in encouraging the data access
measure
Medical trials (Romanini et al. 2020) It is based on machine and profound learning that relies on expository examination to
perceive conceivable clinical preliminary applicants, where scientists can contract
down their pool from a wide assortment of information
Information crowdsourcing (Rodrigues et al. 2016) The wellbeing field has been publicly supporting, and nowadays’ specialists utilize the
strategy to get to a tremendous measure of information that individuals transfer
Outbreak prediction (Chen et al. 2019a, b) Machine and profound learning-based procedures are utilized to screen and expect
flare-ups about the world to anticipate the scourge
Medical imaging diagnostics (Nasser et al. 2019) Simulated intelligence strategies end up being broader, just as productive in their
capacity to see an expanding measure of information sources from different clinical
pictures
13
Fig. 4 Alzheimer’s disease

detection using artificial intelli-
gence techniques (Subasi 2020)
original data is stored. The received data is used for training analytical classification of Alzheimer’s disease. Autoencoder
the model to identify the Alzheimer’s disease symptoms. for feature selection formed accuracy up to 98.8% and 83.7%
As seen in Fig. 4, (Subasi 2020) proposed a broad for guessing conversion from mild cognitive impairment, a
framework for detecting Alzheimer’s illness using AI prodromal stage of Alzheimer’s disease.
methods. The learning process is the process of optimizing A deep neural network uses neuroimaging data without
model parameters using a training dataset or prior prac- pre-processing for feature collection that yields accuracies
tice. Learning models can be predictive, predicting the up to 96.0% for Alzheimer’s disease categorization and
future, descriptive, collecting data from input data sources, 84.2% for the medical council of India conversion problems
and combining them. Two critical stages are performed in (Oomman et al. 2018). Chen et al. (2017) hypothesized the
machine learning and deep learning: pre-processing the combination of diffusivity and kurtosis in diffusion kurto-
vast input and improving the model. The second phase sis imaging to increase the capacity of diffusion kurtosis
involves effectively testing the learning model and resem- imaging in detecting Alzheimer’s disease. The method
bling the answer. Oh et al. (2019) offered a technique for was applied on the 53 subjects, including 27 Alzheimer’s
demonstrating the end-to-end learning of four binary clas- patients, which provides an accuracy of 96.23%. Janghel
sification problems using a volumetric convolutional neu- et al. (2020) used a convolution neural network to improve
ral network form. The trials are performed on the ADNI classification accuracy. They demonstrated a deep learn-
database, and the results indicated that the suggested tech- ing technique for identifying Alzheimer’s disease using
nique obtained an accuracy of 86.60% and a precision of data from the Alzheimer’s disease neuroimaging initiative
73.95%, respectively. Raza et al. (2019) proposed a unique database, which included magnetic resonance imaging and
AI-based examination and observation of Alzheimer’s dis- positron emission tomography scan pictures of Alzheimer’s
order. The analysis results appeared at 82% improvement patients, as well as an image of a healthy individual. The
in contrast with notable existing procedures. experiment attained an average classification accuracy of
Additionally, above 95% precision is accomplished to 99.95% for the magnetic resonance imaging dataset and
order the exercises of everyday living, which are very reas- 73.46% for the positron emission tomography scan dataset.
suring regarding checking the action profile of the subject. Balaji et al. (2020) presented the gait classification system
Lodha et al. (2018) used a machine-learning algorithm based on machine learning to help the clinician diagnose
to process the data obtained by neuroimaging technolo- the stage of Parkinson’s disease. They used four supervised
gies to detect Alzheimer’s in its primitive stage. It uses machine learning algorithms: decision tree, support vector
various algorithms like support vector machine (SVM), machine, ensemble classifier, and Bayes’ classifier, which
gradient boosting, K-nearest neighbour, Random forest, a are used for statistical and kinematic analysis that predict
neural network that shows the accuracy rate 97.56, 97.25, the severity of Parkinson’s disease.
95.00, 97.86, 98.36, respectively. Lei et al. (2020) state that
to evaluate Alzheimer’s ailment, a clinical score forecast 4.2 Diagnosis of cancer disease
using neuroimaging data is incredibly profitable since it can
adequately reveal the sickness status. The proposed struc- Artificial Intelligence methods can affect several facets of
ture comprises three sections: determination dependent on cancer therapy, including drug discovery, drug develop-
joint learning, highlight encoding dependent on profound ment, and the clinical validation of these drugs. Pradhan
polynomial arrange and amass learning for relapse through et al. (2020) evaluated several machine learning algorithms
help vector relapse technique. Jo et al. (2019) performed which are flexible for lung cancer recognition correlated
the deep learning approach and neuroimaging data for the with the internet of things. They reviewed various papers
13
to predict different diseases using a machine learning algo- alternations in gastrointestinal and liver cancer. Also, these
rithm. They also identified and depicted various research methods could be used as an inexpensive biomarker only if
directions based on the existing methodologies. Memon the pathology workflows are used. Kohlberger et al. (2019)
et al. (2019) proposed an AI calculation-based symptomatic built up a convolution neural organization to restrict and
framework which adequately grouped the threatening and measure the seriousness of out-of-fold districts on digitized
favorable individuals in the climate of the internet of things. slides. On contrasting it and pathologist-reviewed center
They tried the proposed strategy on the Wisconsin Diagnos- quality, ConvFocus accomplished Spearman rank coeffi-
tic Breast Cancer. They exhibited that the recursive element cients of 0.81 and 0.94 on two scanners and replicated the
determination calculation chose the best subset of highlights typical designs from stack checking. Tschandl et al. (2019)
and the classifier support vector machine that accomplished build an image-based artificial intelligence for skin cancer
high order precision of 99% and affectability 98%, and diagnosis to address the effects of varied representations
Matthew’s coefficient is 99%. Das et al. (2019) proposed of clinical expertise and multiple clinical workflows. They
another framework called the watershed Gaussian-based also found that excellent quality artificial intelligence-based
profound learning method to depict the malignant growth clinical decision-making support improved diagnostic accu-
injury in processed tomography pictures of the liver. They racy over earlier artificial intelligence or physicians. It is
took a test of 225 pictures which are used to build up the observed that the least experienced clinicians gain the most
proposed model. Yue et al. (2018) reviewed the machine from AI-based support. Chambi et al. (2019) worked on the
learning techniques that include artificial neural networks, volumetric Optical coherence tomography datasets acquired
support vector machines, decision trees, and k-nearest neigh- from resected cerebrum tissue example of 21 patients with
bor for disease diagnosis. The author has investigated the glioma tumours of various stages. They were marked as
breast cancer-related applications and applied them to the either non-destructive or limo-invaded based on histopathol-
Wisconsin breast cancer database. Han et al. (2020) focused ogy assessment of the tissue examples. Unlabelled Optical
on the research and user-friendly design of an intelligent coherence tomography pictures from the other nine patients
recommendation model for cancer patients’ rehabilitation were utilized as the approval dataset to evaluate the strategy
schemes. Their prediction also achieved up to 92%. Rod- discovery execution. Chen et al. (2019a, b) proposed a cost-
rigues et al. (2020) proposed utilizing the move learning effective technique, i.e., ARM (augmented reality micro-
approach and profound learning approach in an IoT frame- scope), that overlays artificial intelligence-based information
work to help the specialists analyse common skin sores, onto the current view of the model in real-time, enabling a
average nevi, and melanoma. This investigation utilized two flawless combination of artificial intelligence into routine
datasets: the first gave by the International Skin Imaging workflows. They even anticipated that the segmented reality
Collaboration at the worldwide Biomedical Imaging Sym- microscope would remove the barrier to using AI considered
posium. The DenseNet201 extraction model, joined with the to enhance the accuracy and efficiency of cancer analysis.
K nearest neighbor classifier, accomplished an exactness of
96.805% for the International Society for Bioluminescence
and Chemiluminescence - International Standard Industrial 4.3 Diabetes detection
Classification dataset. Huang et al. (2020) reviewed the lit-
erature on the application of artificial intelligence for can- Diabetes Mellitus, also known as diabetes, is the leading
cer diagnosis and prognosis and demonstrated how these cause of high blood sugar. AI is cost-effective to reduce
methods were advancing the field. Kather et al. (2019) used the ophthalmic complications and preventable blindness
deep learning to mine clinically helpful information from associated with diabetes. This section covers the study of
histology. It can also predict the survival and molecular various researchers that worked on detecting diabetes in
Fig. 5 Blood glucose prediction

approaches (Woldaregy et al.
2019)
13
patients (Chaki et al. 2020). Kaur and Kumari (2018) used training and testing of artificial neural networks. Lukmanto
machine learning models on Pima Indian diabetes dataset et al. (2015) worked on many diabetes mellitus patients to
to see patterns with risk factors with the help of the R data provide an advantage for researchers to fight against it. Their
manipulation tool. They also analyse d five predictive mod- main objective was to leverage fuzzy support vector machine
els using the R data manipulation tool and support vector and F-score feature selection to classify and detect diabetes
machine learning algorithm: linear kernel support vector mellitus. The methodology is applied to the Pima Indian
machine, multifactor dimensionality reduction, and radial Diabetes dataset, where they got an accuracy of 89.02% to
basis function. predict the diabetes mellitus patients. Wang et al. (2017)
As shown in Fig. 5, blood glucose prediction has been proposed a weighted rank support vector machine to over-
categorized in three different parts: physiology-based, infor- come the imbalanced problem seen during the daily dose
mation-driven, and hybrid-based. Woldaregy et al. (2019) system of drugs, leading to poor prediction results. They
developed a compact guide in machine learning and a hybrid also employed the area under the curve (AUC) to show the
system that focused on predicting the blood glucose level in model’s effectiveness and improved the average precision
type 1 diabetes. They mentioned various machine learning of their proposed algorithm. Carter et al. (2018) showcased
methods crucial to regulating an artificial pancreas, deci- the performance of 46 different machine learning models
sion support system, blood glucose alarm applications. They compared on re-sampled trained and tested data. The model
had also portrayed the knowledge about the blood glucose obtained the area under the curve of 0.73 of training data and
predictor that gave information to track and predict blood 0.90 of tested data. Nazir et al. (2019) proposed a technique
glucose levels as many factors could affect the blood glucose to minutely detect the diabetic retinopathy’s different stages
levels like BMI, stress, illness, medications, amount of sleep, via tetragonal local octa pattern features that are further clas-
etc. Thus blood glucose prediction provides the forecasting sified by extreme machine learning. For classifying peri-
of an individual’s blood glucose level based on the past and odic heart rate variability signals and diabetes, Swapna et al.
current history of the patient to give an alarm to delay any (2018) presented a deep learning architecture. The authors
complications. Chaki et al. (2020) provided detailed infor- used long short term memory, a convolution neural network,
mation to detect diabetes mellitus and self-management to extract the dynamic features of heart rate variability. They
techniques to prove its importance to the scientists that work achieved an accuracy of 95.7% on using electrocardiography
in this area. They also analyse d and diagnosed diabetes signals along with the support vector machine classification.
mellitus via its dataset, pre-processing techniques, feature
extraction methods, machine learning algorithms, classifica- 4.4 Diagnose chronic diseases
tion, etc. Mercaldo et al. (2017) proposed a method to clas-
sify diabetes-affected patients using a set of characteristics Researchers have shown that artificial intelligence helps in
selected by a world health organization and obtained the pre- the streamlining care of chronic diseases. Therefore, vari-
cision value and recall value 0.770 and 0.775, respectively, ous machine learning algorithms are developed to identify
with the help of the Hoeffding tree algorithm. Mujumdar patients at higher risk of chronic disease. The other tech-
et al. (2019) proposed the model for prediction, classifica- niques based on AI are stated below (Jain et al. 2018).
tion of diabetes, and external factors like glucose, body mass Jain et al. (2018) presented a survey to showcase feature
index, insulin, age, etc. They also analyse d that classifica- choice and arrangement methods to analyse and anticipate
tion accuracy proved to be much more efficient with the the constant illnesses. They utilized dimensionality decrease
new dataset than their used dataset. Kavakiotis et al. (2017) strategies to improve the presentation of AI calculation. To
conducted a systematic review regarding the machine learn- put it plainly, they introduced different component deter-
ing applications, data mining techniques, and tools used in mination techniques and their inalienable points of interest
the diabetes field to showcase the prediction and diagnosis and impediments. He et al. (2019) proposed a kernel-based
of diabetes, its complications, and genetic conditions and structure for training the chronic illness detector to fore-
situation, including the physical condition care management. cast and track the disease’s progression. Their approach was
After the in-depth search, it had been found that supervised based on an enhanced version of a structured output sup-
learning methods characterized 85%, and the rest, 15%, were port vector machine for longitudinal data processing. Tang
characterized by unsupervised learning methods. Aggarwal et al. (2020) utilized deep residual networks to identify
et al. (2020) demonstrated the non-linear heart rate variabil- chronic obstructive pulmonary disease automatically. After
ity in the prediction of diabetes using an artificial neural net- gathering data from the PanCad project, which includes ex-
work and support vector machine. The author computed 526 smokers and current smokers at high risk of lung cancer, the
datasets and obtained the classification accuracy of 90.5% residual network was trained to diagnose chronic obstructive
with a support vector machine. Besides that, they evaluated pulmonary disease using computed topography scans. Addi-
thirteen non-linear heart rate variability parameters for the tionally, they ran three rounds of cross-validation on it. With
13
the help of three-fold cross-validation, the experiment had kidney disease dataset and performed ten-fold cross-valida-
an area under the curve of 0.889. Ma et al. (2020) proposed tion testing. The dataset had been pre-processed for com-
the heterogeneous changed artificial neural organization to pleting and normalizing the missing data. They achieved
identify, divide, and determine persistent renal disappoint- the detection accuracy of 99% and were further tested using
ment utilizing the web of medical things stage. The proposed four patient data samples to predict the disease. Kim et al.
strategy was named uphold vector machine and multilayer (2019) indicated the constant sicknesses in singular patients
perceptron alongside the back engendering calculation. that utilized a character repetitive neural organization to
They used ultrasound images and later performed segmen- regard the information in each class as a word, mainly when
tation in that image. Especially in Kidney segmentation, it an enormous bit of its information esteem is absent. They
performed very well by achieving high results. Aldhyani applied the Char-recurrent neural network to characterize the
et al. (2020) proposed the system that was used to increase Korea National Health and Nutrition Examination Survey
the accuracy in detecting chronic disease by using machine cases. They indicated the aftereffects of higher precision for
learning algorithms. The machine learning methods such the Char-recurrent neural network than for the customary
as Naïve Bayes, support vector machine, K nearest neigh- multilayer perceptron model. Ani et al. (2017) proposed a
bour, and random forest were presented and compared. patient monitoring system for stroke-affected people that
They also used a rough k-means algorithm to figure out the reduced future recurrence by alarming the doctor and pro-
ambiguity in chronic disease to improve its performance. vided the data analytics and decision-making based on the
The Naïve Bayes method and RKM achieved an accuracy patient’s real-time health parameters. That helped the doc-
of 80.55% for diabetic disease, the support vector machine tors in systematic diagnosis followed by tailored treatment
achieved 100% accuracy for kidney disease, and the support of the disease.
vector machine achieved 97.53% for cancer disease. Chui
and Alhalabi (2017) reviewed the chronic disease diagnosis 4.5 Heart disease diagnosis
in smart health care. They provide a summarized view of
optimization algorithms and machine learning algorithms. Researchers suggest that artificial intelligence can predict
The authors also gave information regarding Alzheimer’s the possible periods of death for heart disease patients. Thus
disease, dementia, tuberculosis, etc., followed by the chal- multiple algorithms have been used to predict the heart rate
lenges during the deployment phase of the disease diagno- severity along with its diagnosis. Escamila et al. (2019)
sis. Nam et al. (2019) introduced the internet of things and proposed a dimensionality decrease strategy to discover the
digital biomarkers and their relationships to artificial intel- highlights of coronary illness utilizing the highlight determi-
ligence and other current trends. They have also discussed nation procedure. The dataset used was the UCIrvine artifi-
the role of artificial intelligence in the internet of things for cial intelligence vault called coronary illness which contains
chronic disease detection. Battineni et al. (2020) reviewed 74 highlights. The most remarkable precision was accom-
the applications of predictive models of machine learning to plished by the chi-square and head segment investigation
diagnose chronic disease. After going through 453 papers, alongside the irregular woods classifier. Tuli et al. (2019)
they selected only 22 studies from where it was concluded proposed a Health fog framework to integrate deep learning
that there were no standard methods that would determine in edge computing devices and incorporate it into the real-
the best approach in real-time clinical practice. The com- life application of heart detecting disease. They consisted of
monly used algorithms were support vector machine, logistic the hardware and software components, including body area
regression, etc. Wang et al. (2018) analyse d chronic kidney sensor network, gateway, fogbus module, data filtering, pre-
disease using machine learning techniques based on chronic processing, resource manager, deep learning module, and
Fig. 6 Cardiovascular health

promotion and disease preven-
tion (George et al. 2018)
13
ensembling module. The health fog model was an internet and Decision tree to detect normal and irregular heartbeat
of things-based fog enabled model that can help effectively sounds. Also, after trying, it was discovered that pre-pro-
manage the data of heart patients and diagnose it to identify cessing approach increased the performance of classifying
the heart rate severity. algorithms. The devices utilized for IoT set up were the
George et al. (2018) aimed to describe the obstacles LM35 sensor, Pulse sensor, AD8232 electrocardiographic
Indian nurses face in becoming active and valued members sensor, and Arduino Uno. Thai et al. (2017) proposed a new
of the cardiovascular healthcare team as cardiovascular lightweight method to remove the noise from electrocardio-
disease imposed substantial and increasing physical, psy- graphic signals to perform minute diagnosis and prediction.
chological, societal, and financial burdens. As shown in Initially, they worked on the Sequential Recursive algorithm
Fig. 6, there are numerous possible facts for health intel- for the transformation of signals into digital format. The
ligent mediations to support helping cardiovascular health same was sent to the Discrete Wavelet Transform algorithm
and decreasing hazard for cardiovascular disease. So the to detect the peaks in the data for removing the noises. Then
focus has started on the inhibition of cardiovascular disease features were extracted from the electrocardiographic data-
and, more importantly, on the advancement of cardiovascu- set from Massachusetts Institute of Technology-Beth Israel
lar health. Several findings revealed that depression is con- Hospital to perform diagnosis and prediction and remove
nected with inferior cardiovascular health between adults the redundant features using Fishers Linear Discriminant.
without cardiovascular disease. Nashif et al. (2018) proposed a cloud-based heart disease
Haq et al. (2018) created a system based on machine prediction system for detecting heart disease using machine
learning to diagnose the cardiac disease guess using its learning models derived from Java Based Open Access
dataset and worked on seven prominent feature learning- Data Mining Platform, Waikato Environment for Knowl-
based algorithms. It was also observed that the machine edge Analysis. They got an accuracy level of 97.53% using a
learning-based decision support system assisted the doc- support vector machine with 97.50% sensitivity and 94.94%
tors in diagnosing the heart patients effectively. Khan and specificity. They used an efficient software tool that trained
Member (2020) proposed a framework to estimate the cardio the large dataset and compared multiple machine learning
disease using a customized deep convolution network for techniques. The smartphone used to detect and predict heart
categorizing the fetched sensor information into the usual disease based on the information acquired from the patients.
and unusual state. Their results demonstrated that if there Hardware components are used to monitor the system con-
would be the utmost amount of records, the multi-task cas- tinuously. Babu et al. (2019) aimed to determine whether
caded convolution neural network achieved an accuracy of the heart attack could occur using hereditary or not. Thus to
98.2%. Ahmed (2017) explained the architecture for heart work on it, initially, they collected and compared the previ-
rate and other techniques to understand using machine learn- ous data of parents with their child dataset to find the pre-
ing algorithms such as K nearest neighbour classification to diction and accurate values. It could help them to determine
predict the heart attack during collecting heart rate datasets. how healthy the child is. The authors used different param-
The author also mentioned the six data types predicting heart eters to show the dependent and independent parameters to
attack in three different levels (Patel 2016). The dataset used find whether the person gets a heart attack.
consists of 303 instances and 76 attributes. They worked on
a technique that could reduce the number of deaths from 4.6 Tuberculosis disease detection
heart diseases. They compared various decision tree algo-
rithms to present the heart disease diagnosis using Waikato AI is placed as an answer for aid in the battle against
Environment for Knowledge Analysis. They aimed to fetch tuberculosis. Computerized reasoning applications in
the hidden patterns by using data mining techniques linked indicative radiology might have the option to give pre-
to heart disease to predict its presence. Saranya et al. (2019) cise methods for recognizing the infections for low pay
proposed a cloud-based approach based on sensors for an countries. Romero et al. (2020) performed the classi-
automated disease predictive system to calculate various fication tree analysis to reveal the associations between
parameters of patients like blood pressure, heartbeat rate, predictors of tuberculosis in England. They worked on
and temperature. As per their knowledge, this method could the American Public Health Association data ranging
reduce the time complexity of the doctor and patient in pro- from demographic herd properties and tuberculosis vari-
viding medical treatment quickly. The best part was that ables using Sam Tuberculosis management. They used
anyone could access it from anywhere. Isravel et al. (2020) a machine-learning algorithm, performed data prepara-
presented a pre-processing approach that might enhance the tion, data reduction, and data analysis, and finally got the
accuracy in identifying the electrocardiographic signals. results. Horvath et al. (2020) performed the automatic
They evaluated the classification using different classify- scanning and analysis on 531 slides of tuberculosis, out
ing algorithms such as K nearest neighbour, Naïve Bayes, of which 56 were from the positive specimen. They also
13
validated a scanning and analysis system to combine overall severity probability, separate probabilities of both
fully automated microscopy using deep learning analysis. high severity and low severity forces. Singh et al. (2020)
Their proposed system achieved the highest sensitivity by worked to discover tuberculosis sores in the lungs. They
detecting 40 out of 56 positive slides. Sathitratanachee- proposed a computerized recognition strategy utilizing a
win et al. (2020) developed a convolution neural network profound learning technique known as Antialiased Convo-
model using tuberculosis. They used a specified chest lution Neural Network proposed by Richard Zhang. Their
X-ray dataset taken from the national library of medical dataset included 3D computed topography pictures, which
Shenzhen no. 3 hospitals and did its testing with a non- were cut into 2D pictures. They applied division on each
tuberculosis chest X-ray dataset taken from the national cutting picture utilizing UNet and Link net design.
institute of health care and center. The deep convolution
neural network model achieved the region of curve area 4.7 Stroke and cerebrovascular disease detection
under the curve by 0.9845 and 0.8502 for detecting tuber-
culosis and the specificity 82% and sensitivity of 72%. AI can analyse and detect stroke signs in medical images as
Bahadur et al. (2020) proposed an automatic technique to if the system suspects a stroke in the patient. It immediately
detect the abnormal chest X-ray images that contained at gives the signal to the patient or doctor. Researchers have
least one pathology such as infiltration, fibrosis, pleural proposed various methodologies to showcase the impact
effusion, etc., because of tuberculosis. This technique is of AI in stroke and cerebrovascular detection (Singh et al.
based on a hierarchical structure for extracting the fea- 2009). O’Connell et al. (2017) assessed the diagnostic capa-
ture where feature sets are used in two hierarchy levels bility and temporal stability for the detection of stroke. They
to group healthy and unhealthy people. The authors used observed the mostly identical patterns between the stroke
800 chest X-ray images taken from two public datasets patients and controls across the ten patients. They achieved
named Montgomery and Shenzhen. López-Úbeda et al. the specificity and sensitivity of 90% across the research.
(2020) explored the machine learning methods to detect Labovitz et al. (2017) stated the use of AI for daily monitor-
tuberculosis in Spanish radiology reports. They also ing of patients for the identification and medication. They
mentioned the deep learning classification algorithms achieved the improvement by 50%on plasma drug concentra-
with the purpose of its evaluation and comparison and tion levels. Abedi et al. (2020) also presented a framework
to carry such a task. The authors have used the data of to build up the decision support system using an artificial
5947 radiology reports collected from high-tech media. neural network, which improved patient care and outcome.
Ullah et al. (2020) presented the study of Raman Spec- Singh et al. (2009) compared the different methods to predict
troscopy and machine learning based on principal com- stroke on the cardiovascular health study dataset. They also
ponent analysis and hierarchical component analysis to used the decision tree algorithm for the feature selection
analyse tuberculosis either in positive form or negative process, principal component analysis to reduce the classifi-
form. They also showed Raman results which indicated cation algorithm’s dimension, and a backpropagation neural
the irregularities in the blood composition collected from network. Biswas et al. (2020) introduced an AI-based system
tuberculosis-negative patients. Panicker et al. (2018) for the location and estimation of carotid plaque as carotid
introduced the programmed technique for the location of intima-media thickness for the same and solid atheroscle-
tuberculosis bacilli from tiny smear pictures. They per- rotic carotid divider discovery and plaque estimations.
formed picture binarization and grouping of distinguished
districts utilizing convolution neural organization. They 4.8 Hypertension disease detection
did an assessment utilizing 22 sputum smear minuscule
pictures. The results demonstrated 97.13% review, 78.4% Researchers have found that AI has been able to diagnose
accuracy, 86.76% F-score for predicting tuberculosis. Lai hypertension by taking input data from blood pressure,
et al. (2020) compared the artificial neural network out- demographics, etc. Krittanawong et al. (2018) summarized
comes, support vector machine, and random forest while the review about the recent computer science and medi-
diagnosing anti-tuberculosis drugs on Taipei Medical cal field advancements. They also illustrated the innova-
University Wanfang Hospital patients. They selected the tive approach of artificial intelligence to predict the early
features via univariate risk factor analysis and literature stages of hypertension. They also stated that AI plays
evaluation. The authors achieved the specificity by 90.4% a vital role in investigating the risk factors for hyperten-
and sensitivity of 80%. Gao et al. (2019) investigated the sion. However, on the side, it has also been restricted by
applications of computed topography pulmonary images researchers because of its limitations in designing, etc.
to detect tuberculosis at five levels of severity. They pro- Arsalan et al. (2019) conducted the experiments using three
posed a deep Res Net to predict the severity scores and publicly available datasets as digitized retinal imagery for
analyse the high severity probability. They also calculate vessel extraction (DRIVE), structured analysis of retina
13
Fig. 7 Pulmonary hypertension (Kanegae et al. 2020)
(STARE) for hypertension detection. They achieved the hypertension as artificial intelligence can recognize hyper-
accuracy for all datasets with sensitivity, specificity, area tension’s risk factors and phenotypes.
under the curve, and accuracy of 80.22%, 98.1%, 98.2%, Moreover, it is used to interpret data from randomized
96.55%, respectively. Kanegae et al. (2020) used machine trials that contained blood pressure targets associated with
learning techniques to validate the prediction of risk for new- cardio vascular outcomes. Kiely et al. (2019) investigated the
onset hypertension. They used data in a split form for the prescient model dependent on the medical care assets that
model construction and development and validation to test could be sued to screen huge populaces to distinguish the
its performance. The models they used were XGBoost and patients at great danger of pneumonic blood vessel hyper-
ensemble, in which the XGBoost model was considered the tension. They took the information of 709 patients from
best predictor because it was systolic blood pressure nature 2008 to 2016 with pneumonic blood vessel hypertension
during cardio ankle vascular. Figure 7 shows the structure and contrasted it and separated associate of 2,812,458 who
of heart during its normal phase as well as in hypertension was delegated non-aspiratory blood vessel hypertension just
phase. When the human heart is in hypertension phase, its as the prescient model was created and approved by utiliz-
pulmonary arteries gets constricted because of which the ing cross approval. Kwon et al. (2020) did the past group
right ventricle did not get the blood in to the lungs. learning of information taken on or after successive diseased
Koshimizu et al. (2020) has also described artificial people from two health care sectors to predict pulmonary
intelligence in pulse the executives, which was utilized to hypertension using electrocardiography with the help of
foresee the chance of circulatory strain utilizing enormous artificial intelligence. Sakr et al. (2018) assessed and ana-
scope information. The authors also focused on the meas- lyse d AI strategies, such as Logit Boost, Bayesian Network
ure that was used to control blood pressure using an artifi- Classifier, locally weighted Naïve Bayes, counterfeit neural
cial neural network. In a nutshell, they were trying to prove organization, Support Vector Machine, and Random Tree
that an artificial neural network is beneficial for high blood Forest foresee the people to recognize hypertension. Thus,
pressure organization and can also use it to create medical AI provides insights for hypertension healthcare and imple-
confirmation for the realistic organization of hypertension. ments prescient, customized, and pre-emptive methodolo-
Mueller et al. (2020) stated that using artificial analytic tools gies in clinical practice.
to the large dataset based on hypertension would generate
questionable results and would also miss treatments and the 4.9 Skin disease diagnosis
potential targets. The author also stated that the vision of
hypertension would be challenging to achieve and doubt- Researchers have developed an AI system that can precisely
lessly not happen in the future. Chaikijuraja et al. (2020) group cutaneous skin problems and fill in as an auxiliary
also noted the merits of using artificial intelligence to detect instrument to improve the demonstrative exactness of clini-
cians. Chakraborty et al. (2017) proposed a neural-based
13
location technique for various skin disorders. They utilized information such as type of dataset, techniques, and the pre-
two infected skin pictures named Basel Cell Carcinoma and dicted outcomes regarding the work done by the researchers
Skin Angioma. Non-overwhelming arranging hereditary on different diseases, which in return helped the author to
calculation is used to prepare the counterfeit neural organi- look for the best technique for detecting or diagnosing any
zation, contrasted with the neural network particle swarm particular disease.
optimization classifier and neural network Caesarean Section From Table 4, we can observe that AI techniques have
classifier. Zaar et al. (2020) collected the clinical images proven to be the best for detecting diseases with improved
of skin disease from the department of Dermatology at the results. AI uses machine and deep learning models that work
Sahlgrenska University, where artificial intelligence algo- upon training and testing data sets so that the system can see
rithms had been used for the classification, thereby achieving the disease and diagnose it early. In the AI-based model, we
the diagnosis accuracy by 56.4% for the top five suggested initially need to train human beings to remember the data
diseases. Kumar et al. (2019) used a dual-stage approach that and provide accurate results. However, it also deals with
combined computer vision and machine learning to evalu- the problem. Suppose the training data produced the incor-
ate and recognize skin diseases. During training and testing rect analysis of disease because of insufficient information,
of the diseases, the method produced an accuracy of up to which artificial intelligence cannot factor. As a result, it will
95%. Kolkur et al. (2018) developed a system that identi- become a horrible condition for the patients as AI cannot
fied skin disease based on input symptoms. They collected assure us whether the prediction regarding disease detection
the data of the symptoms of ten skin diseases and got 90% is accurate.
above accuracy. On assaying the accuracy of algorithms in diagnosing
the disease, deep learning classifiers have dominated over
4.10 Liver disease detection machine learning models in the field of disease diagnosis.
Deep learning models have proved to be best in terms of
Researchers have found that AI can treat liver disease at its scalp disease by 99%, Alzheimer disease by 96%, thyroid
early diagnosis to work on its endurance and heal rate. Abdar disease by 99%, 96% in skin disease, 99.37% in case of
et al. (2018) showed that efficient early liver disease recogni- Arrhythmia disease, 95.7% in diabetic disease, while as
tion through Multilayer Perceptron Neural Network calcula- machine learning models achieved 89% in diabetic disease,
tion depends on different choice tree calculations, such as 88.67% in tuberculosis, 86.84% in Alzheimer disease, etc.
chi-square programmed communication indicator and char-
acterization, and relapse tree with boosting strategy. Their
technique had the option to analyse and characterize the liver
malady proficiently. Khaled et al. (2018) introduced an artifi- 5 Discussion
cial neural network for the diagnosis of hepatitis virus. Pro-
tein and Histology is utilized as an info variable for the fake We have presented recently published research studies that
neural organization model, and it also showed the correct employed AI-based Learning techniques for diagnosing the
prediction of diagnosis by 93%. Spann et al. (2020) provided disease in the current review. This study highlights research
the strengths of machine learning tools and their potential as on disease diagnosis prediction and predicting the post-oper-
machine learning is applied to liver disease research, includ- ative life expectancy of diseased patients using AI-based
ing clinical, molecular, demographic, pathological, and radi- learning techniques.
ological data. Nahar and Ara (2018) explored the early guess Investigation 1: Why do we need AI?
of liver ailment using various decision tree techniques. The We know that AI is the simulation of human processes
choice tree methods utilized were J48, Licensed Massage by machines (computer systems) and that this simulation
Therapist, Random Forest, Random Tree, REP tree, Deci- includes learning, reasoning, and self-correction. We require
sion Stump, and Hoeffding Trees. Their primary purpose AI since the amount of labour we must perform is rising
was to calculate and compare the performances of various daily. As a result, it’s a good idea to automate regular tasks.
decision tree techniques. Farokhzad et al. (2016) used fuzzy It conserves the organization’s staff and also boosts produc-
logic for diagnosing liver sickness. Using this method, where tion (Vasal et al. 2020).
they had two triangular membership and Gussy membership In terms of the healthcare industry, AI in health refers to a
functions, they reached 79–83% accuracy. set of diverse technologies that enable robots to detect, com-
prehend, act, and learn1 to execute administrative and clini-
4.11 Comparative analysis cal healthcare activities. AI has the potential to transform
healthcare by addressing some of the industry’s most press-
In addition to the above mentioned reported work, the com- ing issues. For example, AI can result in improved patient
parative analysis illustrated in Table 4 showcase the detailed outcomes and increased productivity and efficiency in care
13
Table 4 Comparative analysis for different disease detection

Authors Type of disease Dataset Technique Reported outcomes
Naseer et al. (2019) Skin disease Primary Tumor data col- Multi-Layer Perceptron Accuracy: 76.67%
lected from Institute of (MLP), Artificial Neural
Oncology Network
Chuang et al. (2011) Liver disease Real time data collected CBR, BPNN, Logistic Accuracy: 95%
from patients Regression, Classifica- Sensitivity: 98%
tion Specificity: 94%
Musleh et al. (2019) Liver disease Data collected from 583 ANN model Accuracy: 99%
liver patients
Chen et al. Urology disease Urology disease related Cox Regression, Machine 71.8% concluded that
(2019a, b) heterogeneous dataset learning, Neural Net- artificial intelligence is
work, Decision support superior in diagnosis of
system urology disease detection
Plawaik et al. (2018) Arrhythmia disease MIT-BIH arrhythmia Deep genetic ensemble of Sensitivity: 94.62%
database classifiers, ECG signal Accuracy: 99.37%
Specificity: 99.66%
Nithya et al. (2020) Kidney disease Kidney ultrasound images ANN, Kmeans clustering, Accuracy: 99.61%
Linear and quadratic
based segmentation
Owasis et al. (2019) Gastrointestinal disease Endoscopic videos with Residual Network, LSTM Area under Curve: 97.057%
52,471 frames
Luo et al. Gastrointestinal cancer Images from Sun Yat-sen GRAIDS, Clopper Pearson Accuracy : 95%
(2019) University cancer centre Method
Khan et al. (2020) Gastrointestinal disease Data collected from VGG 16, ANN, Deep Accuracy: 98.4%
humans through IoT Learning
Gouda et al. (2020) COVID-19 disease CT scan dataset Artificial Intelligence Sensitivity: 90.9%
Specificity: 87.5%
Vasal et al. (2020) COVID-19 disease Chest X-ray dataset Deep Learning models, Accuracy 98.8%
VGG16, DenseNet121,
ResNet50
Minaee et al. (2020) Covid-19 disease 5000 Chest X-ray dataset CNN, ResNet 18, ResNet Sensitivity: 97%
50, Squeeze Net, Specificity: 90%
DenseNet121
Arsalan et al. (2019) Hypertension disease DRIVE, CHASE-DB1, Vess-net Method, AI, Sensitivity: 80.22%
STARE Semantic Segmentation Specificity: 98.1%
Accuracy: 96.55%
Kanegae et al. (2020) Hypertension disease 18,258 patients data col- XGBoost, ensemble,, logis- AUC of
lected from 2005 to 2016 tic regression XGBoost: 0.877
Ensemble: 0.881
Logistic Regression: 0.859
Kiely et al. (2019) Pulmonary Arterial Hyper- Data collected from Hos- Gradient Boosting tree Specificity: 99.99%
tension pital Episode Statistical algorithm
population
Kaur and Kumari (2018) Diabetic disease Pima Indian Diabetes SVM, Radial Basis Accuracy of
dataset Function, KNN, ANN, SVM: 0.89
multifactor, dimensional- KNN: 0.88
ity reduction ANN: 0.86
MDR: 0.83
Lukmanto et al. (2015) Diabetic disease Pima Indian Diabetes Fuzzy support vector Accuracy: 89.02%
dataset machine, SVM
Swapna et al. (2018) Diabetic disease Real time data collected SVM,CNN, Long Short Accuracy: 95.7%
from 20 diabetic and 10 Term Memory
normal people
Lai et al. (2020) Tuberculosis Data taken from Taipei ANN, Random Forest Accuracy: 88.67%
Medical university. Sensitivity: 80%
Specificity: 90.4%
Gao et al. Tuberculosis 100 CT TB images Deep Learning, ResNet Accuracy: 85.29%
(2019)
13
Table 4 (continued)
Panicker et al. (2018) Tuberculosis 22 sputum smear micro- CNN, Image Processing Recall: 97.13%
scopic images Precision: 78.4%
F-score: 86.76%
Rajalakshmi et al. (2018) Retinopathy disease Retinal Images of 296 AI software Sensitivity: 95%
patients Specificity: 80.2%
Keenan et al. (2020) Retinal Fluid detection 1127 SDOCT scan data AI software tool Accuracy: 0.805
Sensitivity: 0.468
Specificity: 0.970
Sarao et al. (2020) Retinopathy detection Real time data of 165 Image Analysis Software, Sensitivity: 90.8%
patients AI software tool Specificity: 75.3%
Shkolyar et al. (2019) Bladder Tumor detection Data of 95 patients from CystoNet, deep learning Sensitivity: 90.9%
TURBT Specificity: 98.6%
Naser and Naseer (2019) Tumor detection Primary Tumor taken from Multilayer Perceptron, Accuracy: 76.67%
Institute of Oncology ANN
Ljubic et al. (2020) Alzheimer’s disease detec- EMR dataset LSTM, RNN, deep learn- AUC : 0.98-0.99
tion SCRP dataset ing model
Khan et al. (2020) Alzheimer’s disease OASIS database Machine learning, Pipeline, Accuracy: 86.84%
Pattern Recognition
Janghel et al. (2020) Alzheimer’s disease ADNI database SVM, KNN, Decision Tree Accuracy: 73.46%
Ahmed (2017) Cardiac Arrest ANFIS dataset Machine learning, KNN, Accuracy: 96%
IoT
Isravel et al. (2020) Heart disease Health dataset KNN, Naïve Bayes, Deci- Accuracy: 80%
sion Tree, ECG signals Sensitivity: 60%
Nashif et al. (2018) Cardiovascular disease Open Access heart disease Data Mining, Machine Accuracy: 97.53%
prediction dataset Learning, SVM, WEKA Specificity: 94.94%
Sensitivity: 97.50%
Bibault et al. (2020) Chronic obstructive pul- ECLIPSE dataset Artificial Intelligence AUC: 0.886
monary disease software tool
Battineni et al. (2020) Chronic disease 22 studies from CINHAL SVM, Logistic Regression Accuracy: 73.1–91.6%
dataset
Aldhyani et al. (2020) Chronic disease Chronic disease dataset SVM, KNN,NB, Random Accuracy: 80.55%
Forest, Sensitivity: 80.14%
Specificity: 80.14%
Precision: 90%
F-score 84.78%
Rodrigues et al. (2020) Skin Lesion ISIC dataset CNN, VGG Net, KNN, Accuracy: 96.805%
Support Vector Machine,
Random Forest
Das et al. Liver cancer 255 Medical images Gaussian Mixture Model, Accuracy: 99.38%
(2019) DNN classifier
Memon et al. (2019) Breast cancer Wisconsin Diagnostic SVM, Machine Learning , Accuracy: 99%
Breast Cancer Sensitivity: 98%
Specificity: 99%
Romanini et al. (2020) Oral cancer Real data collected from ANN, Fuzzy logic Accuracy: 78.89%
dental clinic
Fukuda et al. (2019) Vertical root fracture 330 VRF teeth CNN, DetectNet Precision: 0.93
Recall: 0.75
F measure: 0.83
Chui et al. (2020) Oral cancer 408 OSCC patients KNN, Decision Tree, Accuracy: 70.59%
Support Vector Machine, Sensitivity: 41.98%
Logistic Regression, Specificity: 84.12%
Principal Component
Analysis
Rodrigues et al. (2016) Large Artery Occlusion 750 CTA based dataset LVO algorithm, Artificial Sensitivity: 92%
detection stroke Intelligence Specificity: 90%
13
Table 4 (continued)
Chatterjee et al. (2018) Cerebrovascular large ves- 650 CTA based dataset Large Vessel Occlusion Specificity: 94%
sel detection algorithm, artificial intel- Sensitivity: 82%
ligence
Nazir et.al Diabetic Retinopathy Large scale DR-datasets Content Based Image Accuracy: 99.6%
(2019) detection Retrieval Precision: 0.991
Recall: 0.9932
AUC: 0.995
Ani et al. (2017) Chronic disease detection 191 stroke and non-stroke Random forest, Naïve Accuracy: 93%
patients Bayes, KNN, Classifica-
tion
Bhatt et al. (2019) Thyroid disease Data taken from pregnant Artificial Neural Network, Accuracy: 98.22%
ladies Random forest, Multiple
Regression
Hosseinzadeh et al. (2020) Thyroid disease MRI based dataset Artificial Neural Network Accuracy: 99%
Oh et al. (2019) Alzheimer’s disease ADNI database Convolution Neural Accuracy: 86.60%
Network
Ostovar et al. (2020) Covid 19 disease RTPCR laboratory based Deep learning, Health Specificity: 60–70%
dataset Technology Assessment
Yadav et al. (2020) Thyroid disease 3710 thyroid patients Decision Tree, Random Accuracy of
forest, classification, Decision tree: 98%
regression tree Random forest: 99%
Tengnah et al. (2019) Hypertension Real time dataset Fuzzy logic, Multi-Layer Sensitivity: 90.48%
Perceptron, Support Vec- Specificity: 71.79%
tor Machine, Decision Predicitively: 81.48%
Tree
Tang et al. Pulmonary disease PanCan dataset Deep learning, deep AUC: 0.886
(2020) residual network
Jo et.al (2019) Alzheimer’s disease AD based dataset Recurrent Neural Network, Accuracy: 96.0%
Convolution Neural
Network
Damiani et al. (2020) Squamous Cell Carcinoma Scalp cSCC patients data Artificial Neural Network Accuracy: 91.7%
Sensitivity: 97.6%
Specificity: 85.7%
Morabito et al. (2016) Scalp disease AD and EEG based data Deep Learning, Convolu- Accuracy: 80%
tion neural network,
Multi-Layer Perceptron
Chang et al. (2018) Scalp disease Data collected from scalp Deep learning, Recurrent Precision: 97.41–99.09%
hair physiotherapist Neural Network
addition, it may potentially help bring life-saving medicines

to market more quickly. Figure 8 shows the significance of
AI in the medical field.
Investigation 2: Why is AI important, and how is it used
to analyse the disease?
The emergence of new diseases remains a critical param-
eter in human health and society. Hence, the advances in AI
allow for rapid processing and analysis of such massive and
complex data. It recommends the correct decision for over
ten different diseases (as mentioned in the literature) with
Fig. 8 Importance of artificial intelligence in healthcare
at least 98% accuracy.
Doctors use technologies such as computed tomography
delivery (Gouda et al. 2020). It can also enhance healthcare scan or magnetic resonance imaging to produce a detailed
practitioners’ daily lives by spending more time caring for 3D map of the area that needs to be diagnosed. Later, AI
patients, therefore increasing staff morale and retention. In technology analyse s the system-generated image using
13
machine and deep learning models to spot the diseased Therefore, AI models for disease detection have an ample
area’s features in seconds. As shown in the framework sec- opportunity to drive earlier diagnosis for patients in need
tion, an artificial intelligence model using machine and deep and guide pharmaceutical companies with highly advanced,
learning algorithms is initially trained with the help of a targeted diagnostics to help these patients get correctly diag-
particular disease dataset (Owasis et al. 2019). The dataset nosed and treated earlier in their disease journey (Keenan
is then pre-processed using data cleaning and transforma- et al. 2020). The research work mentioned in the literature
tion techniques so that the disease symptoms in the form has covered both machine and deep learning models for
of feature vectors can be extracted and further diagnosed. diagnosing the diseases such as cancer, diabetes, chronic,
Suppose doctors do not use AI techniques. In that case, heart disease, alzheimer, stroke and cerebrovascular, hyper-
it will cause a delay in treating the patients as it is tough tension, skin, and liver disease. Machine learning models,
to interpret the scanned image manually, and it also takes Random Forest Classifier, Logistic Regression, Fuzzy log-
a considerable amount of time. But, on the other hand, it ics, Gradient Boosting Machines, Decision Tree, K nearest
shows that an AI technique helps the patients and helps the neighbour (KNN), and Support vector machines (SVM) are
doctors save the patient’s life by treating them as early as primarily used in literature. Among deep learning models,
possible (Luo et al. 2019). Convolutional Neural Networks (CNN) have been used most
Investigation 3: What is the impact of AI in medical commonly for disease diagnosis. In addition, faster Recur-
diagnosis? rent Convolution Neural Network, Multilayer Perceptron,
Due to advancements in computer power, learning algo- Long Short Term Memory (LSTM) have also been used
rithms, and the availability of massive datasets (big data) extensively in the literature. Figure 10 displays the usage of
derived from medical records and wearable health monitors. AI-based prediction models in the literature.
The best part of implementing AI in healthcare is that it Investigation 5: What are the challenges faced by the
helps to enhance various areas, including illness detection, researchers while using AI models in several disease
disease classification, decision-making processes, giving diagnosis?
optimal treatment choices, and ultimately, helping people Although AI-based techniques have marked their signifi-
live longer. In terms of disease diagnosis, AI has been used cance in disease diagnosis, there are still many challenges
to enhance medical diagnosis (Chen et al. 2019a, b). For faced by the researchers that need to be addressed.
example, the technology, which is currently in use in China,
may detect hazardous tumors and nodules in patients with i. Limited Data size The most common challenge faced
lung cancer, allowing physicians to provide an early diag- by most of the studies was insufficient data to train the
nosis rather than sending tissue samples to a lab for testing, model. A small sample size implies a smaller training
allowing for earlier treatment (Keenan et al. 2020). Figure 9 set which does not authenticate the efficiency of the
illustrates the influence of artificial intelligence and other proposed approaches. On the other hand, good sample
approaches. size can train the model better than the limited one
Investigation 4: Which AI-based algorithm is used in dis- (Rajalakshmi et al. 2018).
ease diagnosis? ii. High dimensionality Another data-related issue faced
Disease detection algorithms driven by AI demon- in cancer research is high dimensionality. High dimen-
strated to be an effective tool for identifying undiagnosed sionality is referred to a vast number of features as
patients with under-diagnosed, uncoded, and rare diseases. compared to cases. However, multiple dimensionality
Fig. 9 Comparison between AI

and other techniques
13
Fig. 10 Artificial intelligence-based prediction models
reduction techniques are available to deal with this decision-making while diagnosing and treating patients’ ill-
issue (Bibault et al. 2020). nesses using artificial intelligence-based approaches. Not only
iii. Efficient feature selection technique Many stud- that, AI assists physicians in detecting diseases by utilizing
ies have achieved exceptional prediction outcomes. complicated algorithms, hundreds of biomarkers, imaging
However, a computationally effective feature selec- findings from millions of patients, aggregated published clini-
tion method is required to eradicate the data cleaning cal studies, and thousands of physicians’ notes to improve the
procedures while generating high disease prediction accuracy of diagnosis.
accuracy (Koshimizu et al. 2020).
iv. Model Generalizability A shift in research towards
improving the generalizability of the model is 6 Conclusion and future scope
required. Most of the studies have proposed a predic-
tion model that is validated on a single site. There is When it comes to disease diagnosis, accuracy is critical for
a need to validate the models on multiple sites that planning, effective treatment and ensuring the well-being of
can help improve the model’s generalizability (Fukuda patients. AI is a vast and diverse realm of data, algorithms,
et al. 2019). analytics, deep learning, neural networks, and insights that is
v. Clinical Implementation AI-based models have constantly expanding and adapting to the needs of the health-
proved their dominance in medical research; still, the care industry and its patients. According to the findings of this
practical implementation of the models in the clinics study, AI approaches in the healthcare system, particularly for
is not incorporated. These models need to be validated illness detection, are essential. Aiming at illuminating how
in a clinical setting to assist the medical practitioner in machine and deep learning techniques work in various disease
affirming the diagnosis verdicts (Huang et al. 2020). diagnosis areas, the current study has been divided into several
sections that cover the diagnosis of alzheimer’s, cancer, diabe-
Investigation 6: How artificial intelligence-based tech- tes, chronic diseases, heart disease, stroke and cerebrovascu-
niques are helping doctors in diagnosing diseases? lar disease, hypertension, skin disease, and liver disease. The
AI improves the lives of patients, physicians, and hospital introduction and contribution were covered in the first section,
managers by doing activities usually performed by people but in followed by an evaluation of the quality of the work and an
a fraction of the time and the expense. For example, AI assists examination of AI approaches and applications. Later, various
physicians in making suggestions by evaluating vast amounts illness symptoms and diagnostic difficulties, a paradigm for
of healthcare data such as electronic health records, symptom AI in disease detection models, and various AI applications in
data, and physician reports to improve health outcomes and healthcare were discussed. The reported work on multiple dis-
eventually save the patient’s life (Kohlberger et al. 2019). Addi- eases and the comparative analysis of different techniques with
tionally, this data aids in the improvement and acceleration of the used dataset as well as the results of an applied machine and
13
deep learning methods in terms of multiple parameters such as Scalzo F, Li J, Zand R (2020) Using artificial intelligence for
accuracy, sensitivity, specificity, an area under the curve, and improving stroke diagnosis in emergency departments: a practi-
cal framework. Ther Adv Neurol Disord. https://d oi.o rg/1 0.1 177/
F-score has also been portrayed. Finally, the work that assisted 1756286420938962
researchers in determining the most effective method for detect- Aggarwal Y, Das J, Mazumder PM, Kumar R, Sinha RK (2020) Heart
ing illnesses is finished, as in future scope. In a nutshell, medi- rate variability features from nonlinear cardiac dynamics in iden-
cal experts better understand how AI may be used for illness tification of diabetes using artificial neural network and support
vector machine. Integr Med Res. https://doi.org/10.1016/j.bbe.
diagnosis, leading to more appropriate proposals for the future 2020.05.001
development of AI based techniques. Ahmed F (2017) An Internet of Things (IoT) application for predict-
Contrary to considerable advancements over the past several ing the quantity of future heart attack patients. J Comput Appl
years, the area of accurate clinical diagnostics faces numer- 164:36–40. https://doi.org/10.5120/ijca2017913773
Aldhyani THH, Alshebami AS, Alzahrani MY (2020) Soft clustering
ous obstacles that must be resolved and improved constantly to for enhancing the diagnosis of chronic diseases over machine
treat emerging illnesses and diseases effectively. Even health- learning algorithms. J Healthc Eng. https://doi.org/10.1155/
care professionals recognize the barriers that must be overcome 2020/4984967
before sickness may be detected in conjunction with artificial Alfian G, Syafrudin M, Ijaz MF, Syaekhoni MA, Fitriyani NL, Rhee
J (2018) A personalized healthcare monitoring system for dia-
intelligence. Even doctors do not entirely rely on AI-based betic patients by utilizing BLE-based sensors and real-time
approaches at this time since they are unclear of their ability to data processing. Sensors 18(7):2183
anticipate illnesses and associated symptoms. Thus much work Ali M, Tengnah J, Sooklall R (2019) A predictive model for hyper-
is required to train the AI-based systems so that there will be an tension diagnosis using machine learning techniques. Tel-
emed Technol. https://doi.org/10.1016/B978-0-12-816948-3.
increase in the accuracy to predict the methods for diagnosing 00009-X
diseases. Hence, in the future, AI-based research should be con- Ani R, Krishna S, Anju N, Aslam MS, Deepa OS (2017) IoT based
ducted by keeping the flaw mentioned earlier in consideration patient monitoring and diagnostic prediction tool using ensem-
to provide a mutually beneficial relationship between AI and ble classifier. In: 2017 International conference on advances
in computing, communications and informatics (ICACCI), pp
clinicians. In addition to this, a decentralized federated learning 1588–1593. https://doi.org/10.1109/ICACCI.2017.8126068
model should also be applied to create a single training model Ansari S, Shafi I, Ansari A, Ahmad J, Shah S (2011) Diagnosis of
for disease datasets at remote places for the early diagnosis of liver disease induced by hepatitis virus using artificial neural
diseases. network. IEEE Int Multitopic. https://doi.org/10.1109/INMIC.
2011.6151515
Arsalan M, Owasis M, Mahmood T, Cho S, Park K (2019) Aiding the
Acknowledgements This research work was supported by Sejong diagnosis of diabetic and hypertensive retinopathy using artificial
University research fund. Yogesh Kumar and Muhammad Fazal Ijaz intelligence based semantic segmentation. J Clin Med 8:1446.
contributed equally to this work and are first co-authors. https://doi.org/10.3390/jcm8091446
Babu BS, Likhitha V, Narendra I, Harika G (2019) Prediction and
Declarations detection of heart attack using machine learning and internet of
things. J Comput Sci 4:105–108
Conflict of interest The authors declare that they have no conflict Bahadur T, Verma K, Kumar B, Jain D, Singh S (2020) Automatic
of interest. detection of Alzheimer related abnormalities in chest X-ray
images using hierarchical feature extraction scheme. Expert Syst
Ethical approval All procedures performed in studies involving Appl 158:113514. https://doi.org/10.1016/j.eswa.2020.113514
human participants were in accordance with the ethical standards of Balaji E, Brindha D, Balakrishnan R (2020) Supervised machine
the institutional and/or national research committee and with the 1964 learning based gait classification system for early detection and
Helsinki Declaration and its later amendments or comparable ethical stage classification of Parkinson’s disease. Appl Soft Comput J
standards. 94:106494
Battineni G, Sagaro GG, Chinatalapudi N, Amenta F (2020) Appli-
Human and animal rights This article does not contain any studies with cations of machine learning predictive models in the chronic
the animals performed by any of the authors. disease diagnosis. J Personal Med. https://d oi.o rg/1 0.3 390/j pm10
020021
Informed consent Informed consent was obtained from all individual Behera R, Bala P, Dhir A (2019) The emerging role of cognitive com-
participants included in the study. puting in healthcare: a systematic literature review. J Med Inform
129:154–166
Bhatt V, Pal V (2019) An intelligent system for diagnosing thyroid
disease in pregnant ladies through artificial neural network. In:
Conference on advances in engineering science management and
References technology, pp 1–10. https://doi.org/10.2139/ssrn.3382654
Bibault J, Xing L (2020) Screening for chronic obstructive pulmonary
Abdar M, Yen N, Hung J (2018) Improving the diagnosis of liver dis- disease with artificial intelligence. Lancet Digit Health 2:e216–
ease using multilayer perceptron neural network and boosted e217. https://doi.org/10.1016/S2589-7500(20)30076-5
decision tree. J Med Biol Eng 38:953–965. https://doi.org/10. Biswas M, Saba L, Suri H, Lard J, Suri S, Miner M et al (2020) Two
1007/s40846-017-0360-z stage artificial intelligence model for jointly measurement of
Abedi V, Khan A, Chaudhary D, Misra D, Avula V, Mathrawala D, atherosclerotic wall thickness and plaque burden in carotid
Kraus C, Marshall KA, Chaudhary N, Li X, Schirmer CM,
13
ultrasound. Comput Biol Med 123:103847. https://doi.org/10. Escamilla G, Hassani A, Andres E (2019) A comparison of
1016/j.compbiomed.2020.103847 machine learning techniques to predict the risk of heart fail-
Carter JA, Long CS, Smith BP, Smith TL, Donati GL (2018) PT US ure. Mach Learn Paradig 1:9–26. https://d oi.o rg/ 1 0.1 007/
CR. Expert Syst Appl. https://doi.org/10.1016/j.eswa.2018.08. 978-3-030-15628-2_2
002 Farokhzad M, Ebrahimi L (2016) A novel adapter neuro fuzzy infer-
Chaikijurajai T, Laffin L, Tang W (2020) Artificial intelligence and ence system for the diagnosis of liver disease. J Acad Res Com-
hypertension: recent advances and future outlook. Am J Hyper- put Eng 1:61–66
tens 33:967–974. https://doi.org/10.1093/ajh/hpaa102 Fujita S, Hagiwara A, Otuska Y, Hori M, Kumamaru K, Andica C
Chaki J, Ganesh ST, Cidham SK, Theertan SA (2020) Machine learn- et al (2020) Deep learning approach for generating MRA images
ing and artificial intelligence based diabetes mellitus detection from 3D qunatitative synthetic MRI without additional scans.
and self-management: a systematic review. J King Saud Univ Invest Radiol 55:249–256. https://doi.org/10.1097/RLI.00000
Comput Inf Sci. https://doi.org/10.1016/j.jksuci.2020.06.013 00000000628
Chakraborty S, Mali K, Chatterjee S, Banerjee S, Roy K et al (2017) Fukuda M, Inamoto K, Shibata N, Ariji Y, Kutsana S (2019) Evalua-
Detetction of skin disease using metaheurisrtic supported artifi- tion of an artificial system for detecting vertical root fracture on
cial neural networks. In: Industrial automation and elec- panoramic radiography. Oral Radiol 36:1–7
tromechanical engineering conference, pp 224–229. https://doi. Gao XW, James-Reynolds C, Currie E (2019) Analysis of Alzheimer
org/10.1109/IEMECON.2017.8079594 severity levels from CT pulmonary images based on enhanced
Chambi R, Kut C, Jimenez J, Jo J (2019) AI assisted in situ detec- residual deep learning architecture. Healthc Technol. https://doi.
tion of human glioma infiltration using a novel computational org/10.1016/j.neucom.2018.12.086
method for optical coherence tomography. Clin Cancer Res George A, Badagabettu S, Berra K, George L, Kamath V, Thimmappa
25:6329–6338 L (2018) Prevention of cardiovascular disease in India. Clin Prev
Chang W, Chen L, Wang W (2018) Development and experimental Cardiol 7:72–77. https://doi.org/10.4013/JCPC.JCPC_31_17
evaluation of machine learning techniques for an intelligent hairy Gonsalves AH, Singh G, Thabtah F, Mohammad R (2019) Prediction
scalp detection system. Appl Sci 8:853. https://doi.org/10.3390/ of coronary heart disease using machine learning: an experi-
app8060853 mental analysis. ACM Digit Libr. https://doi.org/10.1145/33429
Chatterjee A, Parikh N, Diaz I, Merkler A (2018) Modeling the impact 99.3343015
of inter hospital transfer network design on stroke outcomes in a Gouda W, Yasin R (2020) COVID-19 disease: CT pneumonia analysis
large city. Stroke 49:370–376 prototype by using artificial intelligence, predicting the disease
Chen Y, Sha M, Zhao X, Ma J, Ni H, Gao W, Ming D (2017) Auto- severity. J Radiol Nucl Med 51:196. https://doi.org/10.1186/
mated detection of pathologic white matter alterations in Alzhei- s43055-020-00309-9
mer’s disease using combined diffusivity and kurtosis method. Gupta N, Verma R, Belho E (2019) Bone scan and SPEC/CT scan in
Psychiatry Res Neuroimaging 264:35–45. https://doi.org/10. SAPHO syndrome. J Soc Nucl Med 34:349. https://doi.org/10.
1016/j.pscychresns.2017.04.004 4103/ijnm.IJNM_139_19
Chen J, Remulla D, Nguyen J, Aastha D, Liu Y, Dasgupta P (2019a) Han Y, Han Z, Wu J, Yu Y, Gao S, Hua D, Yang A (2020) Artificial
Current status of artificial intelligence applications in urol- intelligence recommendation system of cancer rehabilitation
ogy and their potential to influence clinical practice. BJU Int scheme based on IoT technology. IEEE Access 8:44924–44935
124:567–577. https://doi.org/10.1111/bju.14852 Haq AU, Li JP, Memon MH, Nazir S, Sun R (2018) A hybrid intel-
Chen P, Gadepalli K, MacDonald R, Liu Y, Dean J (2019) An aug- ligent system framework for the prediction of heart disease using
mented reality microscope with real time artificial intelligence machine learning algorithms. Mob Inf Syst 8:1–21. https://doi.
integration for cancer diagnosis. Nat Med 25:1453–1457 org/10.1155/2018/3860146
Chuang C (2011) Case based reasoning support for liver disease diag- He K, Huang S, Qian X (2019) Early detection and risk assessment
nosis. Artif Intell 53:15–23. https://doi.org/10.1016/j.artmed. for chronic disease with irregular longitudinal data analysis. J
2011.06.002 Biomed Inform 96:103231
Chui KT, Alhalabi W (2017) Disease diagnosis in smart healthcare: Horvath L, Burchkhardt I, Mannsperger S, Last K et al (2020) Machine
innovation. Technol Appl. https://doi.org/10.3390/su9122309 assisted interperation of auramine stains substantially increases
Chui CS, Lee NP, Adeoye J, Thomson P, Choi S-W (2020) Machine through put and senstivity of micrscopic Alzheimer diagnosis.
learning and treatment outcome prediction for oral cancer. J Oral Alzheimer 125:101993. https://doi.org/10.1016/j.tube.2020.
Pathol Med 49:977–985. https://doi.org/10.1111/jop.13089 101993
Connell GCO, Chantler PD, Barr TL (2017) Stroke-associated pattern Hosseinzadeh M, Ahmed O, Ghafour M, Safara F, Ali S, Vo B, Chiang
of gene expression previously identified by machine-learning H (2020) A multiple multi layer perceptron neural network with
is diagnostically robust in an independent patient population. an adaptive learning algorithm for thyroid disease diagnosis in
Genomics Data 14:47–52. https://doi.org/10.1016/j.gdata.2017. the internet of medical things. J Supercomput. https://d oi.o rg/1 0.
08.006 1007/s11227-020-03404-w
Dabowsa N, Amaitik N, Maatuk A, Shadi A (2017) A hybrid intelligent Huang S, Yang J, Fong S, Zhao F (2020) Artificial intelligence in can-
system for skin disease diagnosis. In: Conference on engineering cer diagnosis and prognosis. Cancer Lett 471:61–71
and technology, pp 1–6. https://doi.org/10.1109/ICEngTechnol. Ijaz MF, Alfian G, Syafrudin M, Rhee J (2018) Hybrid prediction
2017.8308157 model for type 2 diabetes and hypertension using DBSCAN-
Damiani G, Grossi E, Berti E, Conic R, Radhakrishna U, Linder D, based outlier detection, synthetic minority over sampling tech-
Bragazzi N, Pacifico A, Piccino R (2020) Artificial neural net- nique (SMOTE), and random forest. Appl Sci 8(8):1325
work allow response prediction in squamous cell carcinoma Ijaz MF, Attique M, Son Y (2020) Data-driven cervical cancer predic-
of the scalp treated with radio therapy. J Eur Acad Dermatol tion model with outlier detection and over-sampling methods.
Venerel 34:1369–1373. https://doi.org/10.1111/jdv.16210 Sensors 20(10):2809
Das A, Acharya UR, Panda SS, Sabut S (2019) Deep learning based Isravel DP, Silas SVPD (2020) Improved heart disease diagnostic
liver cancer detection using watershed transform and Gaussian IoT model using machine learning techniques. Neuroscience
mixture model techniques. Cogn Syst Res 54:165–175. https:// 9:4442–4446
doi.org/10.1016/j.cogsys.2018.12.009
13
Jain D, Singh V (2018) Feature selection and classification systems for Koshimizu H, Kojima H, Okuno Y (2020) Future possibilities for
chronic disease prediction: a review. Egypt Inform J 19:179–189 artificial intelligence in the practical management of hyperten-
Janghel RR, Rathore YK (2020) Deep convolution neural network sion. Hypertens Res 43:1327–1337. https://doi.org/10.1038/
based system for early diagnosis of Alzheimer’s disease. Irbm s41440-020-0498-x
1:1–10. https://doi.org/10.1016/j.irbm.2020.06.006 Krittanawong C, Bomback A, Baber U, Bangalore S, Tang M, Mes-
Jo T, Nho K, Saykin AJ (2019) Deep learning in Alzheimer’s disease: serli F (2018) Future direction for using artificial intelligence to
diagnostic classification and prognostic prediction using neu- predict and manage hypertension. Curr Hypertens Rep 20:75.
roimaging data. Front Aging Neurosci. https://doi.org/10.3389/ https://doi.org/10.1007/s11906-018-0875-x
fnagi.2019.00220 Kumar Y (2020) Recent advancement of machine learning and deep
Kanegae H, Suzuki K, Fukatani K, Ito T, Kairo K, Beng N (2020) learning in the field of healthcare system. Computational intel-
Highly precise risk prediction model for new onset hyperten- ligence for machine learning and healthcare informatics. De
sion using artificial neural network techniques. J Clin Hypertens Gruyter, pp 7–98
22:445–450. https://doi.org/10.1111/jch.13759 Kumar Y, Singla R (2021) Federated learning systems for healthcare:
Kasasbeh A, Christensen S, Parsons M, Lansberg M, Albers G, perspective and recent progress. In: Rehman MH, Gaber MM
Campbell B (2019) Artificial neural network computed tomog- (eds) Studies in computational intelligence, vol965. Springer,
raphy perfusion prediction of ischemic core. Stroke 50:1578– Cham. https://doi.org/10.1007/978-3-030-70604-3_6
1581. https://doi.org/10.1161/STROKEAHA.118.022649 Kumar A, Pal S, Kumar S (2019) Classification of skin disease using
Katharine E, Oikonomou E, Williams M, Desai M (2019) A novel ensemble data mining techniques. Asia Pac J Cancer Prev
machine learning derived radiotranscriptomic signature of 20:1887–1894. https://doi.org/10.31557/APJCP.2019.20.6.1887
perivascular fat improves cardiac risk prediction using coro- Kumar Y, Sood K, Kaul S, Vasuja R (2020). Big data analytics in
nary CT angiography. Eur Heart J 40:3529–3543. https://doi. healthcare. Springer, Cham, pp 3–21
org/10.1093/eurheartj/ehz592 Kwon J, Jeon H, Kim H, Lim S, Choi R (2020) Comapring the per-
Kather J, Pearson A, Halama N, Krause J, Boor P (2019) Deep learn- formance of artificial intelligence and conventional diagnosis
ing microsatellite instability directly from histology in gastro- criteria for detetcting left ventricular hypertrophy using elec-
intestinal cancer. Nat Med 25:1054–1056 tropcardiography. EP Europace 22:412–419. https://doi.org/10.
Kaur H, Kumari V (2018) Predictive modelling and analytics for dia- 1093/europace/euz324
betes using a machine learning approach. Appl Comput Inform. Labovitz D, Shafner L, Gil M, Hanina A, Virmani D (2017) Using
https://doi.org/10.1016/j.aci.2018.12.004 artificial intelligence reduce the risk of non adherence in patients
Kavakiotis I, Tsave O, Salifoglou A, Maglaveras N, Vlahavas I, on anticoagulation theraphy. Stroke 48:1416–1419. https://doi.
Chouvarda I (2017) Machine learning and data mining meth- org/10.1161/STROKEAHA.116.016281
ods in diabetes research. Comput Struct Biotechnol J15:104– Lai N, Shen W, Lee C, Chang J, Hsu M et al (2020) Comparison of the
116. https://doi.org/10.1016/j.csbj.2016.12.005 predictive outcomes for anti-Alzheimer drug-induced hepatotox-
Keenan T, Clemons T, Domalpally A, Elman M, Havilio M, Agron icity by different machine learning techniques. Comput Methods
E, Chew E, Benyamini G (2020) Intelligence detection versus Programs Biomed 188:105307. https://doi.org/10.1016/j.cmpb.
artificial intelligence detection of retinal fluid from OCT: age- 2019.105307
related eye disease study 2: 10 year follow on study. Ophthal- Lei B, Yang M, Yang P, Zhou F, Hou W, Zou W, Li X, Wang T, Xiao
mology. https://doi.org/10.1016/j.ophtha.2020.06.038 X, Wang S (2020) Deep and joint learning of longitudinal data
Khaled E, Naseer S, Metwally N (2018) Diagnosis of hepatititus for Alzheimer’s disease prediction. Pattern Recognit 102:107247
virus using arificial neural network. J Acad Pedagog Res 2:1–7 Lin L, Shenghui Z, Aiguo W, Chen H (2019) A new machine learn-
Khan MA, Member S (2020) An IoT framework for heart disease ing method for Alzheimer’s disease. Simul Model Pract Theory.
prediction based on MDCNN classifier. IEEE Access 8:34717– https://doi.org/10.1016/j.simpat.2019.102023
34727. https://doi.org/10.1109/ACCESS.2020.2974687 Ljubic B, Roychoudhury S, Cao XH, Pavlovski M, Obradovic S, Nair
Khan A, Zubair S (2020) An improved multi-modal based machine R, Glass L, Obradovic Z (2020) Influence of medical domain
learning approach for the prognosis of Alzheimer’s disease. knowledge on deep learning for Alzheimer’s disease prediction.
J King Saud Univ Comput Inf Sci. https://doi.org/10.1016/j. Comput Methods Programs Biomed. https://doi.org/10.1016/j.
jksuci.2020.04.004 cmpb.2020.105765
Khan A, Khan M, Ahmed F, Mittal M, Goyal L, Hemanth D, Satapa- Lodha P, Talele A, Degaonkar K (2018) Diagnosis of Alzheimer’s
thy S (2020) Gastrointestinal diseases segmentation and classi- disease using machine learning. In: Proceedings—2018 4th inter-
fication based on duo-deep architectures. Pattern Recognit Lett national conference on computing, communication control and
131:193–204. https://doi.org/10.1016/j.patrec.2019.12.024 automation, ICCUBEA, pp 1–4
Kiely DG, Doyle O, Drage E, Jenner H, Salvatelli V, Daniels FA, López-Úbeda P, Díaz-Galiano MC, Martín-Noguerol T, Ureña-López
Rigg J, Schmitt C, Samyshkin Y, Lawrie A, Bergemann R A, Martín-Valdivia M-T, Lunab A (2020) Detection of unex-
(2019) Utilising artificial intelligence to determine patients at pected findings in radiology reports: a comparative study of
risk of a rare disease: idiopathic pulmonary arterial hyperten- machine learning approaches. Expert Syst Appl. https://doi.org/
sion. Pulm Circ 9:1–9. https://doi.org/10.1177/2045894019 10.1016/j.eswa.2020.113647
890549 Lukwanto R, Irwansyah E (2015) The early detection of diabetes melli-
Kim C, Son Y, Youm S (2019) Chronic disease prediction using char- tus using fuzzy hierarchical model. Proc Comput Sci 59:312–319
acter-recurrent neural network in the presence of missing infor- Luo H, Xu G, Li C, Wu Q et al (2019) Real time artificial intelli-
mation. Appl Sci 9:2170. https://doi.org/10.3390/app9102170 gence for detection of upper gastrointestinal cancer by endos-
Kohlberger T, Norouzi M, Smith J, Peng L, Hipp J (2019) Artificial copy: a multicentre, case control, diagnostic study. Lancet Oncol
intelligence based breast cancer nodal metastasis detection. Arch 20:1645–1654. https://doi.org/10.1016/S1470-2045(19)30637-0
Pathol Lab Med 143:859–868 Ma F, Sun T, Liu L, Jing H (2020) Detection and diagnosis of chronic
Kolkur MS, Kalbande DR, Kharkar V (2018) Machine learning kidney disease using deep learning-based heterogeneous
approaches to multi-class human skin disease Ddetection. Innov modified artificial neural network. Future Gener Comput Syst
Healthc Tech 14:29–39 111:17–26
13
Matusoka R, Akazawa H, Kodera S (2020) The drawing of the digital Oomman R, Kalmady KS, Rajan J, Sabu MK (2018) Automatic detec-
era in the management of hypertension. Hypertens Res 43:1135– tion of alzheimer bacilli from microscopic sputum smear images
1140. https://d oi.o rg/1 0.1 161/H
YPERT
ENSIO
NAHA.1 20.1 4742 using deep learning methods. Integr Med Res 38:691–699.
Memon M, Li J, Haq A, Memon M (2019) Breast cancer detection https://doi.org/10.1016/j.bbe.2018.05.007
in the Iot health environment using modified recursive feature Ostovar A, Chimeh E, Fakoorfard Z (2020) The diagnostic value of CT
selection. Wirel Commun Mob 2019:19 scans in the process of diagnosing COVID-19 in medical centers.
Mercaldo F, Nardone V, Santone A, Nardone V, Santone A (2017) Dia- Health Technol Assess Act 4:1–7
betes mellitus affected patients classification diagnosis through Owasis M, Arsalan M, Choi J, Mahmood T, Park K (2019) Artificial
machine learning techniques through learning through machine intelligence based classification of multiple gastrointestinal dis-
learning techniques. Proc Comput Sci 112:2519–2528. https:// eases using endoscopy videos for clinical diagnosis. J Clin Med
doi.org/10.1016/j.procs.2017.08.193 8:786. https://doi.org/10.3390/jcm8070986
Minaee S, Kafieh R, Sonka M, Yazdani S, Soufi G (2020) Deep- Panicker RO, Kalmady KS, Rajan J, Sabu MK (2018) Automatic
COVID: predicting covid-19 from chest X-ray images using deep detection of tuberculosis bacilli from microscopic sputum smear
transfer learning. Comput Vis Pattern Recognit 3:1–9. https://d oi. images using deep learning methods. Biocybern Biomed Eng
org/10.1016/j.media.2020.101794 38(3):691–699. https://doi.org/10.1016/j.bbe.2018.05.007
Momin M, Bhagwat N, Dhiwar A, Devekar N (2019) Smart body Park JH, Cho HE, Kim JH, Wall MM, Stern Y, Lim H, Yoo S, Kim
monitoring system using IoT and machine learning. J Adv Res HS, Cha J (2020) Machine learning prediction of incidence of
Electr Electron Instrum Eng Smart Body Syst Using IoT Mach Alzheimer’s disease using large-scale administrative health data.
Learn 1:1–7. https://doi.org/10.15662/IJAREEIE.2019.08050 Npj Digit Med. https://doi.org/10.1038/s41746-020-0256-0
10 Patel SB (2016) Heart disease using machine learning and data minig
Morabito F, Campolo M, Leracitano C, Ebadi J, Bonanno L, Bar- techniques. Health Technol 10:1137–1144
manti A, Desalvo S, Barmanti P, Ieracitano C (2016) Deep Plawiak P, Ozal Y, Tan R, Acharya U (2018) Arrhythmia detection
Convolutional neural Network for classification of mild cogni- using deep convolution neural network with long duration ECG
tive impaired and Alzheimer’s disease patients from scalp EEG signals. Comput Biol Med 102:411–420. https://doi.org/10.
recordings. Res Technol Soc Ind Levaraging Better Tomorrow. 1016/j.compbiomed.2018.09.009
https://doi.org/10.1109/RTSI.2016.7740576 Pradhan K, Chawla P (2020) Medical Internet of things using machine
Mueller FB (2020) AI (Artificial Intelligence) and hypertension learning algorithms for lung cancer detection. J Manag Anal.
research. Telemed Technol 70:1–7. https://doi.org/10.1007/ https://doi.org/10.1080/23270012.2020.1811789
s11906-020-01068-8 Rajalakshmi R, Subashini R, Anjana R, Mohan V (2018) Automated
Mujumdar A, Vaidehi V (2019) Diabetes prediction using machine diabetic retinopathy detection in smartphone-based fundus pho-
learning. Proc Comput Sci 165:292–299. https://doi.org/10. tography using artificial intelligence. Eye 32:1138–1144. https://
1016/j.procs.2020.01.047 doi.org/10.1038/s41433-018-0064-9
Musleh M, Alajrami E, Khalil A, Nasser B, Barhoom A, Naser S Rathod J, Wazhmode V, Sodha A, Bhavathankar P (2018) Diagnosis
(2019) Predicting liver patients using artificial neural network. J of skin diseases using convolutional neural network. In: Second
Acad Inf Syst Res 3:1–11 international conference on electronics, communication and aero-
Nahar N, Ara F (2018) Liver disease detection by using different tech- space technology, pp 1048–1051. https://doi.org/10.1109/
niques. Elsevier 8:1–9. https://doi.org/10.5121/ijdkp.2018.8201 ICECA.2018.8474593
Nam KH, Kim DH, Choi BK, Han IH (2019) Internet of Things, digital Raza M, Awais M, Ellahi W, Aslam N, Nguyen HX, Le-Minh H (2019)
biomarker, and artificial intelligence in spine: current and future Diagnosis and monitoring of Alzheimer’s patients using classical
perspectives. Neurospine 16:705–711. https://doi.org/10.14245/ and deep learning techniques. Expert Syst Appl 136:353–364.
ns.1938388.194 https://doi.org/10.1016/j.eswa.2019.06.038
Naser S, Naseer I (2019) Lung cancer detection using artificial neural Rodrigues J, Matteo A, Ghosh A, Szantho G, Paton J (2016) Com-
network. J Eng Inf Syst 3:17–23 prehensive characterisation of hypertensive heart disease left
Nashif S, Raihan R, Islam R, Imam MH (2018) Heart disease detection ventricular pehnotypes. Heart 20:1671–1679
by using machine learning algorithms and a real-time cardio- Rodrigues DA, Ivo RF, Satapathy SC, Wang S, Hemanth J, Filho PPR
vascular health monitoring system. Healthc Technol 6:854–873. (2020) A new approach for classification skin lesion based on
https://doi.org/10.4236/wjet.2018.64057 transfer learning, deep learning, and IoT system. Pattern Recog-
Nasser I, Naser S et al (2019) Predicting tumor category using artificial nit Lett 136:8–15. https://doi.org/10.1016/j.patrec.2020.05.019
neural network. Eng Inf Technol 3:1–7 Romanini J, Barun L, Martins M, Carrard V (2020) Continuing edu-
Nazir T, Irtaza A, Shabbir Z, Javed A, Akram U, Tariq M (2019) Arti- cation activities improve dentists self efficacy to manage oral
ficial intelligence in medicine diabetic retinopathy detection mucosal lesions and oral cancer. Eur J Dent Educ 25:28–34
through novel tetragonal local octa patterns and extreme learning Romero MP, Chang Y, Brunton LA, Parry J, Prosser A, Upton P, Rees
machines. Artif Intell Med 99:101695. https://doi.org/10.1016/j. E, Tearne O, Arnold M, Stevens K, Drewe JA (2020) Decision
artmed.2019.07.003 tree machine learning applied to bovine alzheimer risk factors to
Nensa F, Demircioglu A, Rischipler C (2019) Artificial intelligence in aid disease control decision making. Prev Vet Med 175:104860.
nuclear medicine. J Nucl Med 60:1–10. https://doi.org/10.2967/ https://doi.org/10.1016/j.prevetmed.2019.104860
jnumed.118.220590 Sabottke C, Spieler B (2020) The effect of image resolution on deep
Nithya A, Ahilan A, Venkatadri N, Ramji D, Palagan A (2020) Kid- learning in radiography. Radiology 2:e190015. https://doi.org/
ney disease detection and segmentation using artificial neural 10.1148/ryai.2019190015
network and multi kernel k-means clustering for ultrasound Sakr S, El Shawi R, Ahmed A, Blaha M et al (2018) Using machine
images. Measurement 149:106952. https://doi.org/10.1016/j. learning on cardiorespiratory fitness data for predicting hyperten-
measurement.2019.106952 sion: the henry ford exercise testing project. PLoS One 13:1–18.
Oh K, Chung YC, Kim KW, Kim WS, Oh IS (2019) Classification https://doi.org/10.1371/journal.pone.0195344
and visualization of Alzheimer’s disease using volumetric con- Santroo A, Clemente F, Baioochi C, Bianchi C, Falciani F, Valente S
volutional neural network and transfer learning. Sci Rep 9:1–16. et al (2019) From near-zero to zero fluoroscopy catheter ablation
https://doi.org/10.1038/s41598-019-54548-6
13
procedures. J Cardiovasc Electrophys 30:2397–2404. https://doi. Thai DT, Minh QT, Phung PH (2017) Toward an IoT-based expert sys-
org/10.1111/jce.14121 tem for heart disease diagnosis. In: Modern artificial intelligence
Saranya E, Maheswaran T (2019) IOT based disease prediction and and cognitive science conference, vol 1964, pp 157–164
diagnosis system for healthcare. Healthc Technol 7:232–237 Tigga NP, Garg S (2020) Prediction of type 2 diabetes using machine
Sarao V, Veritti D, Paolo L (2020) Automated diabetic retinopathy learning prediction of type 2 diabetes using machine learning
detection with two different retinal imaging devices using artifi- classification methods classification methods. Proc Comput Sci
cial intelligence. Graefe’s Arch Clin Exp Opthamol. https://doi. 167:706–716. https://doi.org/10.1016/j.procs.2020.03.336
org/10.1007/s00417-020-04853-y TranX B, Latkin A, Lan H, Ho R, Ho C et al (2019) The current
Sathitratanacheewin S, Sunanta P, Pongpirul K (2020) Heliyon deep research landscap of the application of artificial intelligence in
learning for automated classification of Alzheimer-related chest managing cerebovasclar and heart disease. J Environ Res Public
X-ray: dataset distribution shift limits diagnostic performance health 16:2699. https://doi.org/10.3390/ijerph16152699
generalizability. Heliyon 6:e04614. https://doi.org/10.1016/j. Tschandl P, Nisa B, Cabo H, Kittler H, Zalaudek I (2019) Expert level
heliyon.2020.e04614 diagnosis of non pigmented skin cancer by combined convolution
Shabut AM, Hoque M, Lwin KT, Evans BA, Azah N, Abu-hassan KJ, neural networks. Jama Dermatol 155:58–65
Hossain MA (2018) An intelligent mobile-enabled expert system Tuli S, Basumatary N, Gill SS, Kahani M, Arya RC, Wander GS (2019)
for alzheimer disease diagnosis in real time. Expert Syst Appl HealthFog: an ensemble deep learning based smart healthcare
114:65–77. https://doi.org/10.1016/j.eswa.2018.07.014 system for automatic diagnosis of heart diseases in integrated
Shkolyar E, Jia X, Chnag T, Trivedi D (2019) Augmented bladder IoT and fog computing environments. Future Gener Comput Syst
tumor detection using deep learning. Eur Urol 76:714–718. 104:187–200. https://doi.org/10.1016/j.future.2019.10.043
https://doi.org/10.1016/j.eururo.2019.08.032 Uehera D, Hayashi Y, Seki Y, Kakizaki S, Horiguchi N, Tojima H,
Singh N, Moody A, Leung G, Ravikumar R, Zhan J, Maggissano Yamazaki Y, Sato K, Yasuda K, Yamada M, Uraoka T, Kasama
R, Gladstone D (2009) Moderate carotid artery stenosis: MR K (2018) Non invasive prediction of non alchlolic steatohepatitus
imaging depicted intraplaque hemorrhage predicts risk of cer- in Japanses patiens with morbid obesity by artificial intelligence
ebovascular ischemic events in asymptomatic men. Radiology using rule extraction technology. World J Hepatol 10:934–943.
252:502–508. https://doi.org/10.1148/radiol.2522080792 https://doi.org/10.4254/wjh.v10.i12.934
Singh J, Tripathy A, Garg P, Kumar A (2020) Lung Alzheimer detec- Ullah R, Khan S, Ishtiaq I, Shahzad S, Ali H, Bilal M (2020) Cost
tion using anti-aliased convolutional networks networks. Proc effective and efficient screening of Alzheimer disease with
Comput Sci 173:281–290. https://doi.org/10.1016/j.eswa.2018. Raman spectroscopy and machine learning algorithms. Photo-
07.014 diagn Photodyn Ther 32:101963. https://d oi.o rg/1 0.1 016/j.p dpdt.
Skaane P, Bandos A, Gullien R, Eben E, Ekseth U, Izadi M, Jebsen I, 2020.101963
Gur D (2013) Comparison of digital mammography alone and Uysal G, Ozturk M (2020) Hippocampal atrophy based Alzheimer’s
digital mammography plus tomo-sysnthesis in a population based disease diagnosis via machine learning methods. J Neurosci
screening program. Radiology 267:47–56 Methods 337:1–9. https://doi.org/10.1016/j.jneumeth.2020.
Sloun R, Cohen R, Eldar Y (2019) Deep learning in ultrasound imag- 108669
ing. IEEE 108:11–29. https://doi.org/10.1109/JPROC.2019. Vasal S, Jain S, Verma A (2020) COVID-AI: an artificial intelligence
2932116 system to diagnose COVID 19 disease. J Eng Res Technol 9:1–6
Soundarya S, Sruthi MS, Sathya BS, Kiruthika S, Dhiyaneswaran J Wang Z, Zhang H, Kitai T (2017) Artificial Intelligence in precision
(2020) Early detection of Alzheimer disease using gadolinium cardiovascular medicine. J Am Coll Cardiol 69:2657–2664
material. Mater Today Proc. https://doi.org/10.1016/j.matpr. Wang Z, Chung JW, Jiang X, Cui Y, Wang M, Zheng A (2018)
2020.03.189 Machine learning-based prediction system for chronic kidney
Spann A, Yasodhara A, Kang J, Watt K, Wang B, Bhat M, Goldenberg disease using associative classification technique. Int J Eng Tech-
A (2020) Applying machine learning in liver disease and trans- nol 7:1161–1167. https://doi.org/10.14419/ijet.v7i4.36.25377
plantation: a survey. Hepatology 71:1093–1105. https://doi.org/ Woldargay A, Arsand E, Botsis T, Mamyinka L (2019) Data driven
10.1002/hep.31103 glucose pattern classification and anomalies detection. J Med
Srinivasu PN, SivaSai JG, Ijaz MF, Bhoi AK, Kim W, Kang JJ (2021a) Internet Res 21:e11030
Classification of skin disease using deep learning neural net- Yadav D, Pal S (2020) Prediction of thyroid disease using decision
works with MobileNet V2 and LSTM. Sensors 21(8):2852 tree ensemble method. Hum Intell Syst Integr. https://doi.org/
Srinivasu PN, Ahmed S, Alhumam A, Kumar AB, Ijaz MF (2021) An 10.1007/s42454-020-00006-y
AW-HARIS based automated segmentation of human liver using Yang J, Min B, Kang J (2020) A feasibilty study of LYSO-GAPD
CT images. Comput Mater Contin 69(3):3303–3319 detector for DEXA applications. J Instrum. https://doi.org/10.
Subasi A (2020) Use of artificial intelligence in Alzheimer’s disease 1088/1748-0221/15/05/P05017
detection. AI Precis Health. https://doi.org/10.1016/B978-0-12- Yue W, Wang Z, Chen H, Payne A, Liu X (2018) Machine learning
817133-2.00011-2 with applications in breast cancer diagnosis and prognosis.
Swapna G, Vinayakumar R, Soman KP (2018) Diabetes detection using Designs 2:1–17. https://doi.org/10.3390/designs2020013
deep learning algorithms. ICT Express 4:243–246. https://doi. Zaar O, Larson A, Polesie S, Saleh K, Olives A et al (2020) Evaluation
org/10.1016/j.icte.2018.10.005 of the diagnositic accuracy of an online artificial intelligence
Tang LYW, Coxson HO, Lam S, Leipsic J, Tam RC, Sin DD (2020) application for skin disease diagnosis. Acta Derm Venereol
Articles towards large-scale case-finding: training and validation 100:1–6. https://doi.org/10.2340/00015555-3624
of residual networks for detection of chronic obstructive pulmo- Zebene A, Årsand E, Walderhaug S, Albers D, Mamykina L, Botsis
nary disease using low-dose CT. Lancet Digit Health 2:e259– T, Hartvigsen G (2019) Data-driven modeling and prediction of
e267. https://doi.org/10.1016/S2589-7500(20)30064-9 blood glucose dynamics: Machine learning applications in type 1
Tegunov D, Cramer P (2019) Real-time cryo-electron microscopy data diabetes. Artif Intell Med 98:109–134. https://doi.org/10.1016/j.
preprocessing with warp. Nat Med 16:1146–1152. https://doi. artmed.2019.07.007
org/10.1038/s41592-019-0580-y Zhang R, Simon G, Yu F (2017) Advancing Alzheimer’s research: a
review of big data promises. J Med Inform 106:48–56
13
Zhang F, Zhang T, Tian C, Wu Y, Zhou W, Bi B et al (2019) Radi- Publisher’s Note Springer Nature remains neutral with regard to
ography of direct drive double shell targets with hard X-rays jurisdictional claims in published maps and institutional affiliations.
generated by a short pulse laser. Nucl Fusion. https://doi.org/10.
1088/1741-4326/aafe30
Zhou Z, Yang L, Gao J, Chen X (2019) Structure–relaxivity relation-
ships of magnetic nanoparticles for magnetic resonance imag-
ing. Adv Mater 31:1804567. https://d oi.o rg/1 0.1 002/a dma.2 0180
4567
13

2021 Article 3612

Uploaded by

Copyright:

Available Formats

2021 Article 3612

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

2021 Article 3612

Uploaded by

Copyright:

Available Formats

Journal of Ambient Intelligence and Humanized Computing (2023) 14:8459–8486

Artificial intelligence in disease diagnosis: a systematic literature

Healthcare is shaping up in front of our eyes with advances

Fig. 1 Distribution of published

1.1 Contribution n-space, and qi and pi are the Euclidean vectors starting

2 Materials and methods P(x|c)P(c)

h(t) = f (h(t−1) , xt ;𝜃) (8) 2.1 Quality assessment

Table 1 Inclusion and exclusion parameters

3.1 Framework for AI in disease detection

AI describes the capability of a machine to study the way a

Fig. 3 Framework for disease

Table 2 Medical imaging types with their respective descriptions

of difference investigation; and ideal development and 4.1 Diagnosis of Alzheimer’s disease

Table 3 Healthcare applications and their purpose

Fig. 4 Alzheimer’s disease

Fig. 5 Blood glucose prediction

Fig. 6 Cardiovascular health

Fig. 7 Pulmonary hypertension (Kanegae et al. 2020)

Table 4 Comparative analysis for different disease detection

addition, it may potentially help bring life-saving medicines

Fig. 9 Comparison between AI

Fig. 10 Artificial intelligence-based prediction models

You might also like

1.1 Contribution n-space, and qi and pi are the Euclidean vectors starting

2 Materials and methods P(x|c)P(c)

h(t) = f (h(t−1) , xt ;𝜃) (8) 2.1 Quality assessment

3.1 Framework for AI in disease detection

of difference investigation; and ideal development and 4.1 Diagnosis of Alzheimer’s disease