Artificial Intelligence

CU6051NA

Coursework 1

2019-20 Autumn

Student Name: Kshitiz Basnet

London Met ID: 17031231

College ID: np01cp4a170139

Assignment Due Date: 01-13-2020

Assignment Submission Date: 01-13-2020

I confirm that I understand my coursework needs to be submitted online via Google Classroom under the
relevant module page before the deadline in order for my assignment to be accepted and marked. I am fully
aware that late submissions will be treated as non-submission and a mark of zero.
Contents

1 Introduction ................................................................................................................... 1

1.1 Machine Learning in AI .......................................................................................... 2

1.2 Problem Domain ..................................................................................................... 3

2 Background ................................................................................................................... 4

2.1 Research on Recommendation System .................................................................... 4

2.2 Review of existing solutions on Recommendation Systems ..................................... 8

2.2.1 System 1 .......................................................................................................... 8

2.2.2 System 2 .......................................................................................................... 9

2.2.3 System 3 ........................................................................................................ 10

3 Solution ....................................................................................................................... 11

3.1 Proposed solution .................................................................................................. 11

3.2 AI Algorithm ........................................................................................................ 12

3.3 Pseudo code of solution......................................................................................... 13

3.4 Diagrams .............................................................................................................. 14

3.4.1 Flowchart of the Solution ............................................................................... 14

3.4.2 State Transition Diagram of the Solution ........................................................ 15

4 Conclusion .................................................................................................................. 16

4.1 Work analysis ....................................................................................................... 16

4.2 Application of solution in the real world ............................................................... 17

4.3 Further work ......................................................................................................... 17

5 Bibliography ............................................................................................................... 18

6 Appendix ..................................................................................................................... 21

6.1 History of AI ......................................................................................................... 21

6.2 AI Today............................................................................................................... 21

6.3 Stages of Artificial Intelligence ............................................................................. 22

6.4 Classification in machine Learning ....................................................................... 23

6.5 AI concepts ........................................................................................................... 23

6.6 Differences between Machine learning and Deep learning..................................... 24

6.7 Cosine similarity ................................................................................................... 24

Table of Figures
Figure 1.Flowchart of the Solution ...................................................................................... 14
Figure 2.State Transition diagram of Solution ..................................................................... 15

Table of Tables

Table 1. Artificial Intelligence in today's industrial sector ................................................... 21

CU6051NA Artificial Intelligence

1 Introduction

Computers are used for almost everything in the present day. The computer technology
has helped humans in almost every field, from entertainment to scientific research.
Artificial Intelligence is the branch of science which is another astonishing development
in computer technologies.

“Artificial Intelligence is a branch of science which deals with helping machines find
solutions to complex problems in a more human-like fashion” (Sridhar, 2018).

Artificial Intelligence is rapidly evolving day-by-day. The growth of AI has helped many
industries like technology, education, manufacturing, entertainment etc. The tasks carried
out by AI is reliable and the outcome of the tasks like data can help in growth of industry
in the future.

The list of stats below shows growth of AI in today’s world:

i. According to Adobe, 46% of enterprises were expected to be using AI till the end
of 2019, share of jobs requiring AI has increased by 450% since 2013, investment
in AI startup by increased by six times since 2000.
ii. According to Stanford University, startups in AI has increased by astonishing
1400% since 200.
iii. According to IDC, Robot imports have increased from around 100,000 in 2000 to
roughly 250,000 in 2015.
(Utermohlen, 2018)

Kshitiz Basnet
17031231 1
CU6051NA Artificial Intelligence

1.1 Machine Learning in AI

Machine learning is often regarded as a synonym to AI. But it is an academic discipline

of AI which deals with applying AI for making systems which have capacity of
imitating human learning aptitudes. Machine learning allows systems to learn and
perform task on its own. It is classified into the following types:

i. Supervised Learning: It uses algorithms which learns from training dataset. The
learning stops when the algorithm achieves its goal. It is further categorized into
the following:
a. Classification
b. Regression

ii. Unsupervised Learning: It uses only input data but not a corresponding output
data. The goal of this type of machine learning is to model the underlying
structure of the data to know more about the data. It is further categorized into
the following:
a. Clustering
b. Association

iii. Semi supervised Learning: It uses large amount of input data and few
corresponding output data. This machine learning type falls in between
supervised and unsupervised learning.
(Brownlee, 2016)

iv. Reinforcement Learning: Reinforcement learning is the training of machine

learning models to make a sequence of decisions. The machine can achieve a
goal in a complex environment. Self-driving cars are the best example of
reinforcement learning. The system uses reward or penalty approaches to
perform a task. The goal is to have a maximum reward. (Osiński & Budek, 2018)

The applications of Machine learning are virtual personal assistant, email spam
filtering, online customer support, product recommendations etc.

Kshitiz Basnet
17031231 2
CU6051NA Artificial Intelligence

1.2 Problem Domain

For generations, humans have depicted their stories in dramas and this has influenced
people, their beliefs and their actions. Today, that depiction is made through movies, plays
and dramas.

Movie industry is one of the most influential and successful entertainment industry. “In
2018 alone the global box office revenue reached 41 Billion dollars and it is expected to
increase in coming years, according to Statista, a website which generates statistics on
media and advertisement.” (Watson, 2019) This data shows that people around the globe
love movies and are willing to spend their money watching them.

Movies are available in internet in various sites. Netflix, Fmovies, Bmovies are some of
the leading applications in streaming the movie online. The progress of technological
development has increased the amount of movies in an exponential manner. This
astronomical increase of movies found in the internet today has also impacted the movie
industry as well. A person is able to explore very vast amount of movies in the internet.

Although the internet stores large database of movies, the problems for the user are:

a. Person also lacks the ability to get the relevant movie he or she wants to
watch due to large amount of available movies.
b. The search for relevant content of movies for the user can be very tiresome
for a user as it can require lot of amount of time.
c. Sometimes the user might not find the movie that is he/she wants to watch
after hours of searching.

Kshitiz Basnet
17031231 3
CU6051NA Artificial Intelligence

2 Background
2.1 Research on Recommendation System

“A recommendation/recommender system is the system that is capable of predicting

the future preference of a set of items for a user, and recommend the top items.” (Huang,
2018).

The recommendation system is the effective tool that can be used in the type of problem
mentioned in the problem domain section. A recommendation engine can show the
accurate recommendations to the user where the user can actually view the relevant
type of movies.

Basically, a recommendation system is used to list the items that the user might be
interested in viewing. This helps both the supplier of items, in this case a movie
application to have more users and also the user to get the relevant movies based on
their preference. The two main approaches towards constructing a recommendation
systems are described in detail in the next page.

Kshitiz Basnet
17031231 4
CU6051NA Artificial Intelligence

a. Content based recommendation

Content based recommendation systems use the content or the attributes of the
products that the user likes and according to that content the relevant product is
recommended to the user. The widely used algorithm used in this type of
recommendation engine is TF*IDF algorithm.

The advantages of Content-based filtering are:

i. It produces more reliable result because this method only has to analyze the
items and single user’s profile.
ii. The items are recommended to user on feature level basis. There is no need
to compare user to another unknown user.
iii. The items can be recommended to the user if there are less number of users.

The dis-advantages of content based recommendation are:

i. The recommendation can be imprecise if the content does not contain

enough information in items.
ii. It is limited to the content or attributes of items of the dataset. (Balu, 2019)

Kshitiz Basnet
17031231 5
CU6051NA Artificial Intelligence

b. Collaborative Filtering:

Collaborative filtering uses user’s behavior to suggest relevant items based on the
similar users of the system. There are two types of collaborative filtering:

i. User-User Collaborative Filtering

This type of collaborative filtering compares most similar users and suggests
relevant product to the user.
In this type of collaborative filtering, the data of rating of items purchased
by the user is stored in a vector. The similarity between uses is calculated
using n*n matrix. Similarity is calculated using cosine similarity formula.
Then, recommendation matrix is calculated and value is calculated for all
items that are new for that user in descending order and top items are
recommended to user.

ii. Item-Item Collaborative filtering

This type of collaborative filtering compares similar items rather than the
similar users. The difference in this type of collaborative filtering is that the
recommendation matrix is m*m matrix where m is number of items present
in the system.

The advantages of collaborative filtering are:

a. It can work reliably and flexibly even if the system has large number
of users.
b. It is based on real life activity of user so, the connection between
disparate items can be made.

Kshitiz Basnet
17031231 6
CU6051NA Artificial Intelligence

The disadvantages of collaborative filtering are as follows:

a. The program can get in scalability problems when the number of users
and items gets too high.
b. The history of user cannot tell whether the item is liked by user or not
if the user does not rate an item which results in data sparsely.
c. If there is no user history, there is an anomaly to show which item to
show to user.

iii. Hybrid Recommendation: The hybrid recommendation system uses both

content based and collaborating filtering to make more advanced
recommendation systems

(Doshi, 2019) (Keenan, 2017)

Kshitiz Basnet
17031231 7
CU6051NA Artificial Intelligence

2.2 Review of existing solutions on Recommendation Systems

The existing recommendation systems which were reviewed are described in detailed
below:

2.2.1 System 1

Project Name: Recommender system based on users’ behavior and collaborative

location and tracking

Project Done by:

 Alberto Huertas Celdrana

 Manuel Gil Pereza
 Felix J. Garcıa Clementeb
 Gregorio Martinez Pereza

Approach Followed: User-User Collaborative Filtering and Content based filtering

Features of System

a. It finds similar users so as to consider their ratings when generating new

recommendations in the Collaborative Filtering models.
b. It create users’ profiles and classifies items using the Context based filtering
approaches.
c. It uses the location and the tracking of users in the location-based Context Aware
Filtering approaches.
d. Taking into account that most users do not give ratings to items, Recommendations
are generated automatically from the generators recommenders automatically.
e. The combination of 3 approaches allows to provide accurate recommendations and
decrease problems like cold-start and sparsity.

(Huertas et al., 2015)

Kshitiz Basnet
17031231 8
CU6051NA Artificial Intelligence

2.2.2 System 2

Project Name: Movie Recommender System

Project Done by:

 Prateek Sappadla
 Yash Sadhwani
 Pranit Arora

Approach Followed: Collaborative Filtering and Content based recommendation.

Features of System

a. Keyword for the searching the Movie Plot.

b. It makes profile of the user’s interests using information collected, and
recommend items based on that profile using content based filtering.
c. It groups similar users together and use information about the group to make
recommendations to the user

(Sappadla et al., 2017)

Kshitiz Basnet
17031231 9
CU6051NA Artificial Intelligence

2.2.3 System 3

Project Name: Movie Recommendation System

Project Done by:

 Vishwa Gosalia
 Bhavesh Chatnani
 Aman Gupta
 Pranal Soni

Approach Followed: Collaborative Filtering and Content based recommendation.

Features of System

a. This movie recommendation system recommends movies to a user by evaluating

IMDB ratings.
b. It asks user to input the genre of the movie and the year of movie released. It
segregates the list of movies from the dataset according to the inputs provided by user
and finally displays the list of movies.
c. Users can rate a movie and get the related movies based on the rating.

(Gosalia et al., 2018)

Kshitiz Basnet
17031231 10
CU6051NA Artificial Intelligence

3 Solution
3.1 Proposed solution

The solution for the problem domain described in the introduction is by using machine
learning to create a content based recommender system. The working of the solution
from user’s perspective is explained by an example below:

Suppose a user has to be recommended a movie. The user will be shown the list of all
the movies from the dataset. The system asks the user to input the name of the movie
which is on the dataset. Then the user is asked to rate the movie. Based on that input
the system recommends the movies which are similar to the movie. The user can input
the movies and rate them for better recommendation.

The programming language which will be used to create the system will be python. The
dataset which will be used will be IMDB movie data set.

Flow of solution

The solution is created by doing the works as shown in these steps shown below:

i. The dataset downloaded is cleaned before using it.

ii. A data frame is created by reading from the comma separated values (csv)
file of the dataset.
iii. The features from the dataset are selected which are columns of the data
frame like movie name, genre, cast, director etc. for the program.
iv. The null values in the columns are transferred to empty values.
v. A new column in data frame is created which combines all the features in
step iii.
vi. A count matrix is created from the newly combined column.
vii. The cosine similarity is implemented on count matrix on whole data frame.
viii. The movie input is taken from user.
ix. The index of the input movie form the dataset is taken into account and all
the similar movies are to be generated in ascending order.
x. The user is shown the list of similar movies.

Kshitiz Basnet
17031231 11
CU6051NA Artificial Intelligence

3.2 AI Algorithm

The algorithm used in the solution is K-Nearest Neighbor. This is used to group the array of
items based on the cosine similarity of items

The algorithm of the KNN used in the solution is shown below:

Step 1. Load the dataset

Step 2. Select the value for K.

Step 3. For each point in the dataset

3.1. Calculate the distance between the query example and the current example
from the data.

3.2. Sort the data and distance in ascending order based on distance between
points.

Step 4. Select K top rows from the sorted array.

Step 5. Get the labels of the selected K entries.

Step 6. If regression, return the mean of the K labels

Step 7. If classification, return the mode of the K labels

Step 8. End

(Harrison, 2018)

Kshitiz Basnet
17031231 12
CU6051NA Artificial Intelligence

3.3 Pseudo code of solution

Import pandas as pd

Import numpy as np

import CountVetorizer

import cosine similarity

df=pd.read_csv(“name of dataset”)

features=list[column_name]

FOR EACH feature (features):

df[‘feature’]= df[‘feature’].fill(“”)

function combinefeatures(row):

return row[‘column1’]+ “ ”+ row[‘column2’]….

df[‘combinefeatures’]=df.individual_row

countmatrix=countvectorizor().fit_transform(df[‘combinefeatures’])

cosine_sim=cosine_similarity(countmatrix)

input movie name movie_liked_by_user as

similarmovies=list(cosine_sim(enumerate (cosine_sim [movie_liked_by_user]))

FOR EACH movie (similarmovies):

Print movie

Input movie name

If exit

End

Kshitiz Basnet
17031231 13
CU6051NA Artificial Intelligence

3.4 Diagrams
3.4.1 Flowchart of the Solution

Figure 1.Flowchart of the Solution

Kshitiz Basnet
17031231 14
CU6051NA Artificial Intelligence

3.4.2 State Transition Diagram of the Solution

Figure 2.State Transition diagram of Solution

Kshitiz Basnet
17031231 15
CU6051NA Artificial Intelligence

4 Conclusion

Artificial Intelligence is the innovation of today. It can be used to make solutions which
make our work easier, faster and reliable. The solution that has been chosen is Movie
recommendation system. This solution can be implemented in the solving real world
problems. This project can help both developers who are willing to build similar projects
and the users who want the relevant movies.

4.1 Work analysis

First of all, movie recommendation system as AI topic for the coursework and it was
approved by the module leader. The conceptual solution required research on many AI
topics like machine learning, use of machine learning to create a recommendation system
through various websites, books, journals and academic reports. The machine learning
procedure was chosen create a recommendation engine. Therefore, the research on
Machine learning was a crucial part. It was done by reading blogs, watching videos,
reading reports and reading books related to machine learning. The recommendation
system could be made by using any approaches. For choosing a suitable approach in
creating a recommendation system a lot of research on academic reports, books, journals
etc. was done. Then, academic reports and journals which published the solutions
recommendation engine were read and understood. This made it clear that content based
recommendation system will be the solution to the problem domain. Finally, the report
writing process was carried out. The algorithm, flowchart and pseudo-code of the
program was written at the end of report.

Kshitiz Basnet
17031231 16
CU6051NA Artificial Intelligence

4.2 Application of solution in the real world

The solution has many applications in the real world. In the movie industry, it can have
a significant use as both the movie providers and the users can largely benefit from it as
users get relevant movies and the movie provider can get more users by implementing
this feature. The developers can create similar systems like product recommendation
system, food recommendation system etc. taking the help from this system.

4.3 Further work

The primary work that is need to be done is implementing the solution to make a program
which can recommend the user similar list of movies. It will be done by creating the
supervised machine learning model. This will predict what the movie will be
recommended to the users. Then, training and testing points will be created which will
be used to train and validate the model created above respectively. This will create a very
effective application. If possible, new features will be added to the solution to increase
its usability and reliability.

Kshitiz Basnet
17031231 17
CU6051NA Artificial Intelligence

5 Bibliography

Balu, B. (2019) Content-Based Recommendation System [Online]. Available from:

https://medium.com/towards-artificial-intelligence/content-based-recommender-system-
4db1b3de03e7 [Accessed 8 January 2020].

Brownlee, J. (2016) Supervised and Unsupervised Machine Learning Algorithms [Online].

Available from: https://machinelearningmastery.com/supervised-and-unsupervised-machine-
learning-algorithms/ [Accessed 7 January 2020].

Bullok, M. (2019) Artificial General Intelligence in plain English [Online]. Available from:
https://towardsdatascience.com/artificial-general-intelligence-in-plain-english-e8f6e9a56555
[Accessed 5 January 2020].

Childs, M. (2011) John McCarthy: Computer scientist known as the father of AI [Online].
Available from: https://www.independent.co.uk/news/obituaries/john-mccarthy-computer-
scientist-known-as-the-father-of-ai-6255307.html [Accessed 3 January 2020].

Davis, S. (2017) 28 Artificial Intelligence Terms You Need to Know [Online]. Available from:
https://dzone.com/articles/ai-glossary [Accessed 3 January 2020].

Doshi, S. (2019) Brief on Recommender Systems [Online]. (1) Available from:

https://towardsdatascience.com/brief-on-recommender-systems-b86a1068a4dd [Accessed 10
January 2020].

Gahukar, G. (2018) Classification Algorithms in Machine Learning [Online]. Available from:

https://medium.com/datadriveninvestor/classification-algorithms-in-machine-learning-
85c0ab65ff4 [Accessed 2 January 2020].

Gosalia, V., Chatnani, , Gupta, & Soni,. (2018) Movie Recommendation System. International
Research Journal of Engineering and Technology, 5(2), p.2.

Harrison, O. (2018) Machine Learning Basics with the K-Nearest Neighbors Algorithm
[Online]. (2) Available from: https://towardsdatascience.com/machine-learning-basics-with-
the-k-nearest-neighbors-algorithm-6a6e71d01761 [Accessed 9 January 2020].

Huang, S. (2018) Introduction to Recommender System. Part 1 (Collaborative Filtering,

Singular Value Decomposition) [Online]. Available from:

Kshitiz Basnet
17031231 18
CU6051NA Artificial Intelligence

https://hackernoon.com/introduction-to-recommender-system-part-1-collaborative-filtering-
singular-value-decomposition-44c9659c5e75 [Accessed 8 January 2020].

Huertas, A., Pérez, G., Perez, M. & Clemente, F.J.G. (2015) Design of a Recommender System
Based on Users’ Behavior and Collaborative Location and Tracking. Journal of Computational
Science., 12(10.1016), pp.83-94.

Kanakiya , B. & Gangwar, A. (2020) Introduction to Deep Learning [Online]. Available from:
https://www.geeksforgeeks.org/introduction-deep-learning/ [Accessed 7 January 2020].

Keenan, T. (2017) What Is Collaborative Filtering? [Online]. (1) Available from:

https://www.upwork.com/hiring/data/how-collaborative-filtering-works/ [Accessed 10
January 2020].

Lateef, Z. (2019) Types Of Artificial Intelligence You Should Know [Online]. Available from:
https://www.edureka.co/blog/types-of-artificial-intelligence/ [Accessed 5 January 2020].

Lewis, T. (2014) A Brief History of Artificial Intelligence [Online]. Available from:

https://www.livescience.com/49007-history-of-artificial-intelligence.html [Accessed 03
January 2020].

next. (2020) What is Artificial Intelligence in the Energy Industry ? [Online]. Available from:
https://www.next-kraftwerke.com/knowledge/artificial-intelligence [Accessed 2 January
2020].

Osiński, & Budek,. (2018) What is reinforcement learning? [Online]. Available from:
https://deepsense.ai/what-is-reinforcement-learning-the-complete-guide/ [Accessed 7 January
2020].

Parker, S. (2020) How is AI transforming the Entertainment Industry? [Online]. Available

from: https://readwrite.com/2020/01/10/how-ai-is-transforming-the-entertainment-industry/
[Accessed 11 January 2020].

Prabhakaran, S. (2018) Cosine Similarity – Understanding the math and how it works (with
python codes) [Online]. (1) Available from: https://www.machinelearningplus.com/nlp/cosine-
similarity/ [Accessed 9 January 2020].

Sappadla, P., Sadhwani, & Arora,. (2017) Movie Recommender System. Project Work. New
York: Semantic Scholar New York University.

Kshitiz Basnet
17031231 19
CU6051NA Artificial Intelligence

Schroer, A. (2019) ARTIFICIAL INTELLIGENCE IN CARS POWERS AN AI REVOLUTION

IN THE AUTO INDUSTRY [Online]. Available from: https://builtin.com/artificial-
intelligence/artificial-intelligence-automotive-industry [Accessed 03 January 2020].

Sridhar, G.S. (2018) Introduction to Ai. In G.S. Sridhar, ed. Artificial Intelligence (Way to
future). G.S. Sridhar. pp.0-1.

Techopedia. (2020) Artificial Superintelligence (ASI) [Online]. Available from:

https://www.techopedia.com/definition/31619/artificial-superintelligence-asi [Accessed 5
January 2020].

Utermohlen, K. (2018) 15 Artificial Intelligence (AI) Stats You Need to Know in 2018 [Online].
Available from: https://towardsdatascience.com/15-artificial-intelligence-ai-stats-you-need-
to-know-in-2018-b6c5eac958e5 [Accessed 3 January 2020].

Watson, A. (2019) Global box office revenue from 2005 to 2018 [Online]. Available from:
https://www.statista.com/statistics/271856/global-box-office-revenue/ [Accessed 7 January
2020].

XRMVision. (2020) Artificial intelligence glossary: a few essential concepts [Online].

Available from: https://xrmvision.com/en/blogue/glossaire-de-lintelligence-artificielle-
quelques-concepts-essentiels/ [Accessed 7 January 2020].

Kshitiz Basnet
17031231 20
CU6051NA Artificial Intelligence

6 Appendix

6.1 History of AI

Jhon MacCarthy is known as “Father of Artificial Intelligence (AI)”. Modern Artificial

Intelligence was founded in 1956 in Hanover at a conference at Dartmouth College.
The term “AI” was coined in this conference. The pioneering figures in the Artificial
Intelligence are Alan Turing, Marvin Minsky, Allen Newell, and Herbert A. Simon who
are regarded as “Founding Fathers of AI”. (Lewis, 2014) (Childs, 2011)

6.2 AI Today

Artificial Intelligence is widely used today in businesses, research, entertainment, health

analysis and many more sectors. The table below represents the use of Artificial
intelligence in today’s world.

Sno. Name of the Sectors Example of AI systems

1. Personal Assistance Services Siri (Apple), Echo (Amazon), Google Assistant
(Google) etc.
2. Smart Cars Self-Driving cars of Drive.AI, Tesla, Google,
Nutonomy etc. (Schroer, 2019)
3. Customer Support in Business Chatbots, Autonomous Robots, Customer
Relationship Management (CRM) solutions etc.

4. Entertainment Recommendation systems, advertisement, user

experience personalization etc. (Parker, 2020)
5. Energy Electricity trading, smart power grids, virtual
power plant. (next, 2020)
Table 1. Artificial Intelligence in today's industrial sector

Kshitiz Basnet
17031231 21
CU6051NA Artificial Intelligence

6.3 Stages of Artificial Intelligence

Artificial Intelligence has three stages which are categorized as follows:

i. Artificial Narrow Intelligence: This stage of AI is known as weak AI. The machines of
this stage can perform the tasks which hare simple and defined narrowly. The machine
does not have a thinking ability. The user defines the function and the system uses the
function to do some tasks. The examples are mobile assistant like Alexa, self-driving
cars, robots etc. All the systems made till date are categorized as Weak AI. (Lateef,
2019)
ii. Artificial General Intelligence: This stage is known as strong AI. The machines of this
stage can have the capability of learning anything and performing anything on its own.
The machines of this stage have a capacity of thinking capacity like humans and
working as humans do. These machines are the machines of future. (Bullok, 2019)
iii. Artificial Super Intelligence: This stage can be known as super strong AI. The machines
of this stage can have cognitive ability to surpass humans. These AI machines would
be better than humans in every aspects. They are also the machines of future.
(Techopedia, 2020)

Kshitiz Basnet
17031231 22
CU6051NA Artificial Intelligence

6.4 Classification in machine Learning

“Classification algorithms let machines assign a category to a data point based on

training data.” (Davis, 2017) Classification uses an algorithm to draw conclusions from
the already stored data and then it uses that data to categorize the new data given to the
system. The application if classification are speech recognition, handwriting
recognition, biometric identification, document classification etc. The classifications
can be either binary (two class or 2 possible outcomes) or multi-class(multiple classes
or multiple outcomes).The algorithms which can be used in classification are

i. Naïve Bayes Algorithm

ii. Support Vector Machine
iii. K-Nearest Neignbour

(Gahukar, 2018)

6.5 AI concepts
Artificial Intelligence is a very board term for defining intelligent machines, intelligent
software and intelligent systems. But, every Artificial intelligence is based on the
concepts which are as follows:

i. Machine Learning
ii. Deep Learning
iii. Natural Language Processing
iv. Robotics
v. Computer Vision
vi. Expert Systems

Kshitiz Basnet
17031231 23
CU6051NA Artificial Intelligence

6.6 Differences between Machine learning and Deep learning

The table below differences Machine learning and Deep learning:

Sno. Machine Learning Deep Learning

1. It is just an academic discipline which is not It deals with real world problems by
concerned with real world software. applying machine learning.
2. Machine learning works on comparatively less Deep Learning can work on vary large
amount of data than Deep Learning. amount of data with high performance.
3. Machine learning divides problem into sub- It solves the problem end to end.
tasks to solve them individually.
4. It takes less time to train. It takes longer time to train.
5. It is dependent on low-end machine. It is dependent into high-end machine.

(Kanakiya & Gangwar, 2020) (XRMVision, 2020)

6.7 Cosine similarity

Cosine similarity is a measure of similarity between two non-zero vectors of an inner product
space that measures the cosine of the angle between them. It is derived from Euclidean dot
product formula:

A.B=|A||B| cos(x), where x= angle between points A and B.

Similarity =cos (x)= A.B/|A||B|

(Prabhakaran, 2018)

Kshitiz Basnet
17031231 24