Arti Arya

Followers

Following

Co-authors

Public Views

Uploads

Papers by Arti Arya

An Empirical Study on Privacy-Preserving Swarm Learning for Cataract Detection

Mechanisms and machine science, 2024

Optimal Scheduling of Processing Unit Using Convolutional Neural Network Architecture

Springer eBooks, Sep 1, 2022

Performance Analysis of Bollinger Bands and Long Short-Term Memory(LSTM) models based Strategies on NIFTY50 Companies

2022 12th International Conference on Cloud Computing, Data Science & Engineering (Confluence), Jan 27, 2022

Hedge funds and financial institutions strive to build consistently profitable automated trading ... more Hedge funds and financial institutions strive to build consistently profitable automated trading strategies which can provide higher returns than stock market index baselines. This work aims to beat the stock market baselines with higher returns by building a system capable of evaluating the performance of different automated trading strategies on various different metrics. An automated strategy is a set of rules according to which a computer makes buy/sell decisions in the stock market. This work leverages predictions generated by two strategies namely Bollinger Bands and Long Short-Term Memory to aid in decision making. The LSTM strategy makes use of predictions from 250 LSTM neural networks (5 models per company), while the Bollinger Bands strategy uses close price, simple moving averages and standard deviations to make buy or sell decisions. The strategies’ performance has been evaluated against historic data (backtesting) and continuous real time data (paper trading) of the stocks in the NIFTY50 index being sourced from the stock market. Upon analysis of the backtested data, it was observed that various strategy configurations have beaten the market baselines over different time periods. The results obtained by this work show that the custom strategies proposed have beaten the market baselines in 35.93% (a third) of all time periods back tested. Thus, investing in the proposed custom strategies produces higher returns than investing in the stock market index for the same time periods.

A Machine Learning based Approach to Multiclass Classification of Customer Loyalty using Deep Nets

International Review on Computers and Software, Mar 31, 2017

Identification of customer’s loyalty is one of the most captivating area of today’s growing busin... more Identification of customer’s loyalty is one of the most captivating area of today’s growing business scenario. For any organization, retaining customer is more important than exploring new customers. In this paper, Deep Belief Network (DBN) based approach is implemented for classifying customer loyalties. Training a Deep Belief Network (DBN) is a tedious task but once it trains, the accuracy of classification improves immensely. It also learns from its environment and does not need to be reprogrammed for new situations completely. After training, classifier relies on weight matrices to classify examples. The proposed approach is tested over real as well as sample datasets. The results so acquired are compared with Deep Neural Networks and Support Vector Machine based approaches, which shows Deep Belief Network (DBN) gives accuracy up to 99%.

Natural Language Processing and Big Data: A Strapping Combination

Springer eBooks, 2022

Metaheuristic Approach for Optimizing Supply-Demand Algorithms

2022 IEEE 7th International conference for Convergence in Technology (I2CT), Apr 7, 2022

Big Data Analysis and Machine Learning for Green Computing

Big Data Analysis for Green Computing, 2021

Ensemble of Multimodal Deep Learning Models for Violin Bowing Techniques Classification

Journal of Advances in Information Technology, Dec 31, 2023

Download

Discerning mental illnesses from social media posts using machine and deep learning algorithms

INTERNATIONAL CONFERENCE ON ADVANCES IN COMMUNICATION TECHNOLOGY AND COMPUTER ENGINEERING

Attention Based Evolutionary Approach for Image Classification

2023 15th International Conference on Computer and Automation Engineering (ICCAE)

Optimal Scheduling of Processing Unit Using Convolutional Neural Network Architecture

Springer eBooks, Sep 1, 2022

A Pragmatic Optimal Approach for Detection of Cyber Attacks using Genetic Programming

2020 IEEE 20th International Symposium on Computational Intelligence and Informatics (CINTI)

Cyber-attacks are becoming an increasing threat to people and daily businesses regularly. Attacke... more Cyber-attacks are becoming an increasing threat to people and daily businesses regularly. Attackers have also been evolving their strategies and methods with time. Every attack carried out has the potential to exploit the system on a large scale. Various Artificial Intelligence (AI) algorithms are used to defend such vulnerabilities. This paper analyzes a novel attack and extracts attackers' intrusion scenarios. Evolutionary Computation Techniques have been remarkably used in the field of cybersecurity. This paper particularly discusses the Distributed Denial Of Service (DDoS) attack. The effect of this attack ranges from a disturbance of an elementary service to causing major threats to critical services. In recent times these attacks have become more intricate and carry a significant threat. Therefore, there is a necessity for an intelligent Intrusion Detection System (IDS) to recognize attacks. In this study, work is carried on the latest dataset called Modern DDoS. This paper comprises of comparing the results of six established classification techniques: Random Forest, Naive Bayes, Stochastic Gradient Descent, Decision Trees, Logistic Regression, and K-Nearest Neighbour (KNN) with the proposed Genetic Programming model. The results show that the proposed Genetic Programming model has better accuracy when compared to various existing methods.

Prediction of Material Properties using Crystal Graph Convolutional Neural Networks

2022 7th International Conference on Machine Learning Technologies (ICMLT)

Question-Centric Evaluation of Descriptive Answers using Attention-Based Architecture

2022 12th International Conference on Cloud Computing, Data Science & Engineering (Confluence), 2022

Evaluating descriptive student answers is a long and important process for examiners in an educat... more Evaluating descriptive student answers is a long and important process for examiners in an educational institution. The amount of resources and time that goes into correcting answer scripts is tremendous. To go along with it, anxiety and further implications of exam scores highlight the need for instant yet accurate exam scores for each student. This calls for the necessity to create automatic evaluation systems to reduce the burden on educational institutions. Current systems use various NLP methods to compare similarity between the student’s answer and the expected answer without considering the context in which the question has been asked. In addition to the expected answer, the approach put forward suggests a method for evaluating the answer with the help of the question. This is accomplished using attention networks to extract relevant features in the expected answer and the student’s answer using the question. These features are then passed through a neural network classifier to evaluate the student’s answer. The proposed architecture yielded an accuracy of 95.88% (+ -1) thus proving that knowledge of the question while evaluating a student’s answer leads to a more effective answer evaluation.

Bidirectional Encoder Representation from Transformers (BERT) Variants for Procedural Long-Form Answer Extraction

2022 12th International Conference on Cloud Computing, Data Science & Engineering (Confluence), 2022

Extracting information from large verbose documents is a gruelling task which requires patience a... more Extracting information from large verbose documents is a gruelling task which requires patience and huge amounts of effort. Lengthy documents like Portable Document Formats (PDFs) contain tonnes of information including tables, figures, etc. which makes it hard to retrieve specific pieces of textual content like step-by-step instructions or procedures. To the best of our knowledge, this work is the maiden effort to efficiently extract such long-form procedural answers from the aforementioned information sources. The proposed approach retrieves succinct responses from the relevant PDFs for a given user query using a transformer model embedded with attention mechanism. This model is trained using a self-made dataset namely, Pro-LongQA, consisting of carefully crafted procedure based questions and answers. A comparative study of one of the state-of-the-art transformer models, namely, Bidirectional Encoder Representation from Transformers (BERT) and its different variants such as RoBerta, Albert and DistilBert for the task of long-form question answering is performed. Among which, BERT and RoBerta proved to be the best performing models for this task with an accuracy of 87.2% and 86.4% respectively.

A Metaphorical Study Of Variants Of Recurrent Neural Network Models For A Context Learning Chatbot

2019 4th International Conference on Information Systems and Computer Networks (ISCON), 2019

In this paper, different types of recurrent neural networks such as GRU, bi-directional LSTM and ... more In this paper, different types of recurrent neural networks such as GRU, bi-directional LSTM and a single forward pass LSTM available for question answering systems are explored, so that it can be used in a high level API to be able to create a chatbot interface for a web application. The networks are compared on the Facebook bAbi dataset to test the question answering functionality on 20 different types tasks available in the dataset. The results from existing models were compared and it is observed that the bi-directional LSTM and the GRU performed better in few tasks, however the single forward pass LSTM performed the best for majority of tasks in the dataset.

A Study of Different Disease Detection and Classification Techniques using Deep Learning for Cannabis Plant

by Arti Arya, Kanaad Pathak, Kris Le, Vidyadhar Handragal, and Prakash Hatti

International Journal of Computing and Digital Systems, 2021

In this paper, different models for disease detection and classification are studied for cannabis... more In this paper, different models for disease detection and classification are studied for cannabis plants. Cannabis plants are used for medical and recreational purposes with its recent legalization in some places. Cannabis farmers face problems in cultivation of the crop since it’s susceptible to multiple disorders. With early detection of the disease in the crop it is possible to prevent large waste of yield in the crop. A real dataset is considered for disease detection and classification purposes which is a combination of text and image data and that has been collected over a period of one and a half years (Feb 2018-August 2019). The models used in this study are Fast Region Convolutional Neural Network(F-RCNN), MobileNet Single Shot Multibox Detector(MobileNet-SSD), You Only Look Once(YOLO) and Residual Network-50 Layers (ResNet50). It is found that the MobileNet-SSD provided the best accuracy amongst all the object detection models that are studied and has a lesser training tim...

Download

Hadoop File System with Elastic ReplicationManagement: An Overview

International Journal of Innovative Research in Science, Engineering and Technology, 2014

This paper gives an overview of how Hadoop File System manages massive data as well as handles sm... more This paper gives an overview of how Hadoop File System manages massive data as well as handles small files. As data is exponentially pouring in from all sides in all domains, it has become a necessity to manage and analyze such huge amount of data to extract useful information. This huge amount of data is technically termed as Big Data, which in turn falls under Data Science. Currently a lot of research is going on how to handle such vast pool of data. The Apache Hadoop is a software framework that uses simple programming paradigm to process and analyze large data sets(Big Data) across clusters of computers. The Hadoop Distributed File System(HDFS) is one such technology that manages the Big Data efficiently. In this paper, an insight of ”how HDFS handles big as well as small amount of data” is presented, reviewed and analyzed. As a result, summarized limitations of existing sys- tems are described in the paper along with the future scope of it.

Download

Automatic Fuzzy Classification Tool for Customer Loyalty using Gaussian Membership Function

Data mining and knowledge engineering, 2010

Retaining customers in current era is of utmost importance and crucial for the success of any bus... more Retaining customers in current era is of utmost importance and crucial for the success of any business enterprise. Identifying potential and loyal customers for any enterprise is of great importance. Customer loyalty is one of the major aspects of Customer Relationship Management (CRM). It is very important for an organization to identify its loyal customers, so that the organization can provide better and special services to these customers in order to enhance their business. In this paper, the construction of fuzzy decision tree has been refined by assigning the weights to each attribute, as all the attributes may not affect the customer loyalty equally. This paper proposes an application, which makes use of fuzzy decision trees for classifying customers into various categories. Rather than identifying a customer crisply falling near to the end boundaries, soft boundaries (concept of fuzziness) are used which classifies the customer into different predefined classes of loyalty fro...

A Naive Deep Nets Based Approach for Authenticating Viral Textual Content on Social Media

Social networks in today’s life is more of necessity that helps in reducing the distance and enab... more Social networks in today’s life is more of necessity that helps in reducing the distance and enable people to stay connected. Networking sites are good means to broadcast news too, but quite often than desired these sites are prevalent with lies and frauds, half-truths and facts. The rapid dissemination of such information through social network sites and other online media can have instant and serious ramifications. Fake or deceptive media content and its diffusion through social networks, such as Twitter create an important and challenging problem. The objective of this work is to explore the challenges involved in identifying tweets with unpredictable media content as fake or real. Researchers have used various machine learning techniques and different kinds of classifiers, such as Naive Bayes, Decision Tree, SVM, RNN, DNN, etc. to identify the trustworthy content on Twitter. Some have perceived that how social media users support or deny rumors in breaking news stories but resul...

An Empirical Study on Privacy-Preserving Swarm Learning for Cataract Detection

Mechanisms and machine science, 2024

Optimal Scheduling of Processing Unit Using Convolutional Neural Network Architecture

Springer eBooks, Sep 1, 2022

Performance Analysis of Bollinger Bands and Long Short-Term Memory(LSTM) models based Strategies on NIFTY50 Companies

2022 12th International Conference on Cloud Computing, Data Science & Engineering (Confluence), Jan 27, 2022

A Machine Learning based Approach to Multiclass Classification of Customer Loyalty using Deep Nets

International Review on Computers and Software, Mar 31, 2017

Natural Language Processing and Big Data: A Strapping Combination

Springer eBooks, 2022

Metaheuristic Approach for Optimizing Supply-Demand Algorithms

2022 IEEE 7th International conference for Convergence in Technology (I2CT), Apr 7, 2022

Big Data Analysis and Machine Learning for Green Computing

Big Data Analysis for Green Computing, 2021

Ensemble of Multimodal Deep Learning Models for Violin Bowing Techniques Classification

Journal of Advances in Information Technology, Dec 31, 2023

Download

Discerning mental illnesses from social media posts using machine and deep learning algorithms

INTERNATIONAL CONFERENCE ON ADVANCES IN COMMUNICATION TECHNOLOGY AND COMPUTER ENGINEERING

Attention Based Evolutionary Approach for Image Classification

2023 15th International Conference on Computer and Automation Engineering (ICCAE)

Optimal Scheduling of Processing Unit Using Convolutional Neural Network Architecture

Springer eBooks, Sep 1, 2022

A Pragmatic Optimal Approach for Detection of Cyber Attacks using Genetic Programming

2020 IEEE 20th International Symposium on Computational Intelligence and Informatics (CINTI)

Prediction of Material Properties using Crystal Graph Convolutional Neural Networks

2022 7th International Conference on Machine Learning Technologies (ICMLT)

Question-Centric Evaluation of Descriptive Answers using Attention-Based Architecture

2022 12th International Conference on Cloud Computing, Data Science & Engineering (Confluence), 2022

Bidirectional Encoder Representation from Transformers (BERT) Variants for Procedural Long-Form Answer Extraction

2022 12th International Conference on Cloud Computing, Data Science & Engineering (Confluence), 2022

A Metaphorical Study Of Variants Of Recurrent Neural Network Models For A Context Learning Chatbot

2019 4th International Conference on Information Systems and Computer Networks (ISCON), 2019

A Study of Different Disease Detection and Classification Techniques using Deep Learning for Cannabis Plant

by Arti Arya, Kanaad Pathak, Kris Le, Vidyadhar Handragal, and Prakash Hatti

International Journal of Computing and Digital Systems, 2021

Download

Hadoop File System with Elastic ReplicationManagement: An Overview

International Journal of Innovative Research in Science, Engineering and Technology, 2014

Download

Automatic Fuzzy Classification Tool for Customer Loyalty using Gaussian Membership Function

Data mining and knowledge engineering, 2010

A Naive Deep Nets Based Approach for Authenticating Viral Textual Content on Social Media

Arti Arya

Uploads

Papers by Arti Arya

Log In