		FeedBack \| Contact Us \| Links \| Site Map

Welcome to Journal of Theoretical and Applied Information Technology

Submit Paper / Call for Papers

Journal receives papers in continuous flow and we will consider articles from a wide range of Information Technology disciplines encompassing the most basic research to the most innovative technologies. Please submit your papers electronically to our submission system at http://jatit.org/submit_paper.php in an MSWord, Pdf or compatible format so that they may be evaluated for publication in the upcoming issue. This journal uses a blinded review process; please remember to include all your personal identifiable information in the manuscript before submitting it for review, we will edit the necessary information at our side. Submissions to JATIT should be full research / review papers (properly indicated below main title).

Journal of Theoretical and Applied Information Technology

November 2023 | Vol. 101 No.22

Title:	REINFORCEMENT LEARNING FOR ENERGY EFFICIENT TASK SCHEDULING IN CLOUD
Author:	D MAMATHA RANI, SUPREETHI K P, BIPIN BIHARI JAYASINGH
Abstract:	Cloud infrastructure is widely used by individuals and organizations across the globe due to its scalability, availability and on-demand service provisioning. Consumers and cloud service providers could have Service Level Agreements (SLAs) for mutual benefits. However, honouring SLAs is possible with efficiency in cloud resource allocation, task scheduling and load balancing. Optimization of cloud infrastructure can have huge impact on Quality of Service (QoS) and consumer satisfaction. Efficient task scheduling is one of the approaches to improve cloud infrastructure performance. However, it is a challenging problem in presence of dynamic workloads, heterogeneous resources and dynamism in Virtual Machine (VM) and physical machine’s idle resources. Existing heuristics based approaches suffer from lessened performance due to aforementioned dynamism in the environment. To address this problem, in this paper, we proposed a task scheduling algorithm named Learning based Efficient Task Scheduling (LbETS). This algorithm is based on Deep Reinforcement Learning (DRL) in the form of Deep Q Network (DQN) which has an agent taking feedback from environment in an iterative process converging into ideal task scheduling decision. Our algorithm could improve QoS in terms of energy efficiency, success rate and execution time. Experimental results revealed that LbETS outperforms many existing task scheduling methods due to its learning based approach.
Keywords:	Reinforcement Learning, Task Scheduling, Deep Q Network, Deep Learning, Cloud Computing
Source:	Journal of Theoretical and Applied Information Technology 30^th November 2023 -- Vol. 101. No. 22-- 2023
Full Text

Title:	BRIDGING WEB 4.0 AND EDUCATION 4.0 FOR NEXT GENERATION USER TRAINING IN ERP ADOPTION
Author:	KOH CHEE HONG, ABDUL SAMAD SHIBGHATULLAH, THONG CHEE LING, SAMER MUTHANA SARSAM, SAAD AHMAD QAZI
Abstract:	This study addresses the critical issue of user comprehension and application within the sphere of cloud-based Enterprise Resource Planning (ERP) systems, a recurrent challenge exacerbated by the intricate nature of these systems. To bridge the existing gaps in training methodologies, a novel paradigm that synergizes Web 4.0 and Education 4.0 modules with traditional ERP systems is proposed. This innovative framework ushers in a paradigm shift in ERP adoption strategies, promising a marked enhancement in user interaction and efficiency. Rigorous qualitative evaluations, conducted with expert panels and potential end-users, provided robust validation of the framework's transformative potential in the realm of user training for ERP systems. This pioneering approach not only makes a substantial academic contribution by reframing the perception of ERP systems but also holds a significant practical value in ameliorating the user experience with cloud-based ERP systems. In essence, the adoption of a Web 4.0-oriented approach in user training heralds a revolutionary shift in ERP adoption strategies, setting a solid foundation for future explorations in this domain.
Keywords:	ERP, Web 4.0, Education 4.0, User Training, Qualitative Research, User Comprehension, Cloud-Based Systems, System Usability.
Source:	Journal of Theoretical and Applied Information Technology 30^th November 2023 -- Vol. 101. No. 22-- 2023
Full Text

Title:	INVESTIGATING THE INFLUENCE OF LGBT AND PORNOGRAPHY CONTENT IN NETFLIX AND DISNEY+ HOTSTARS FILMS ON AUDIENCE ACCEPTANCE OF DEVIANT BEHAVIOR: A PERSPECTIVE FROM INDONESIAN THROUGH PERCEPTIONS, FAMILY ROLES, AND FRIENDSHIP ENVIRONMENT
Author:	HENDY HALIM, DIMAS DWIPUTRA, HANAN MURHARDIANSYAH, LA MANI
Abstract:	Perception, family role, and friendship environment are considered as mediating variables, on the audience's response to deviant behavior in Jakarta, which functions as the dependent variable. Deviant behavior encompasses a range of activities that deviate from social norms and can include content related to pornography and LGBT issues. This study aims to provide a comprehensive understanding of the impact of exposure to pornography and LGBT-related content, as well as the contributing factors to the audience's response to such behaviors within the specific context of Jakarta, a city that still holds traditional values. A quantitative approach is employed in this research, utilizing a survey questionnaire as the primary method of data collection. The participant sample is drawn from diverse demographic backgrounds in Jakarta and is selected using non-probability sampling methods. The survey questionnaire is designed to assess the level of film exposure, as well as to measure perception, family role, and friendship environment in relation to deviant behavior.
Keywords:	Film Exposure, Perception, Family Role, Friendship Environment, Societal Acceptance, Deviant Behavior, Jakarta
Source:	Journal of Theoretical and Applied Information Technology 30^th November 2023 -- Vol. 101. No. 22-- 2023
Full Text

Title:	ENHANCING ERROR MINIMIZATION IN MACHINE LEARNING: A NOVEL APPROACH INTEGRATING GRADIENT-BASED AND DICHOTOMY METHODS COMPARED TO GRADIENT DESCENT
Author:	ABDELHAMID OUAZZANI CHAHDI, WIAM SAIDI, KHALID SATORI, RAFIK LASRI, AB-DELLATIF EL ABDERRAHMANI
Abstract:	This research introduces an innovative technique designed to effectively minimize errors in machine lear-ing, with the intention of subsequently applying it to enhance cloud-computing security. Our approach merges gradient-based optimization with the dichotomy method, streamlining the learning process. Its primary objective is the swift identification of the minimum point of a differentiable and convex cost function. To evaluate its efficacy in comparison to the traditional Gradient Descent approach, we apply it to linear regression models and conduct a comprehensive analysis across various dataset sizes and preci-sion settings. Our experiments reveal significant advantages, including reduced execution time and fewer iterations required for convergence. This research contributes to the advancement of optimization tech-niques in machine learning and deep learning, promising potential benefits for practitioners, especially in the context of cloud computing security.
Keywords:	Deep Learning, Dichotomous Search, Gradient Descent, Linear regression, Machine Learning, Optimization
Source:	Journal of Theoretical and Applied Information Technology 30^th November 2023 -- Vol. 101. No. 22-- 2023
Full Text

Title:	APPLICATION OF MACHINE LEARNING ALGORITHMS FOR FORECASTING SALES OF SHRIMP SEEDS
Author:	DEVI ALFITRIA ANGGRAENI, NILO LEGOWO
Abstract:	Indonesia has a great potential for marine and fisheries, particularly in shrimp cultivation. There are many types of shrimp cultivated by Indonesian farmers, but Vannamei shrimp is one of the main types of shrimp in Indonesia, highly demanded both domestically and for export. With such a large demand for shrimp, shrimp farmers need a supplier of shrimp seeds so they can be cultivated into shrimp. PT Prima Akuakultur Lestari, a shrimp seed supplier, who has been in operation for more than 5 years and can produce 130 million shrimp seeds per year. However, this company faced challenges in achieving their sales targets. Sales targets are set based on the sales history of previous months. However, companies are still manually determining their sales targets so that targets are often not achieved. Machine learning is believed to be a method strategy to predict future shrimp sales. Before using Machine Learning, companies must first know which Machine Learning model is suitable for their use. The aim of this research is to test several machine learning models to find a model that is suitable for companies to use in forecasting shrimp seed sales. The research utilized the CRISP-DM method and analyzed shrimp seed sales data from March 2017 to August 2022. Three machine learning models were tested: K-Nearest Neighbor (KNN), Support Vector Regression (SVR), and Neural Network (NN). RapidMiner software was used for data analysis. The results obtained from this research show that the K-Nearest Neighbors model has the highest accuracy value among the three other models tested, with an RMSE value of 6326408.735 and R² 0.215.
Keywords:	Machine Learning, Forecasting, Shimp Seeds, Sales, CRISP-DM
Source:	Journal of Theoretical and Applied Information Technology 30^th November 2023 -- Vol. 101. No. 22-- 2023
Full Text

Title:	A NOVEL APPROACH BASED ON FEATURE SELECTION AND GENE CLASSIFICATION USING SUPPORT VECTOR MACHINES AND QUANTUM ANT LION OPTIMIZATION
Author:	KOMMANA SWATHI, SUBRAHMANYAM KODUKULA
Abstract:	Gene selection for cancer prediction is a crucial model for the medical sector to successfully treat cancer patients. The current model finds it difficult to evaluate the links between the variables used to categorize genes due to the wealth of information available on genes. Local optima traps, slower convergence, and overfitting are drawbacks of the current models. This study suggests using the Quantum Ant Lion (QAL) feature selection optimization to enhance gene classification performance. In the Ant Lion approach, the quantum search process is used to boost search performance, which aids in boosting exploration and avoiding the local optima trap. To maximize utilization of the feature selection based on the fitness function, the Archimedes spiral search is applied in the QAL approach. Exploration and exploitation are increased by the QAL approach, which also serves to raise the method's convergence rate. For classifying genes, the DNN-CNN model and QAL approach both had accuracy of 93.5% and 97.4%, respectively.
Keywords:	Gene Selection, Archimedes Spiral, Quantum Search, Quantum Ant Lion.
Source:	Journal of Theoretical and Applied Information Technology 30^th November 2023 -- Vol. 101. No. 22-- 2023
Full Text

Title:	SECURING AND COMPRESSING TEXT FILES USING THE AES 256 ALGORITHM AND LEVENSTEIN CODE
Author:	HANDRIZAL, FAUZAN NURAHMADI, MUHAMMAD FADLI
Abstract:	In using technology, humans cannot exchange information without the internet. However, along with the development of technology, Data security, and confidentiality are significant issues to be considered in communication since there is a negative impact in the form of data tapping, which results in data being seen and information being taken or possessed by persons who do not have access rights. Methods for securing data are needed in this case. Cryptography is extremely appropriate in the field of data security. The AES algorithm is a security system for protecting data. However, in this case, the impact of cryptography is to create large files, so compression techniques are needed. As a result, a compression method is needed, specifically the Levenstein code, required to compress ciphertext in smaller size. The test used 2 types of characters, namely homogeneous and heterogeneous characters. The test parameters used Ratio of Compression, Compression ratio, space-saving, and bitrate. From the test results, it was found that encryption with AES 256 experienced an increase in the number of characters, but the resulting character length has the same ciphertext length. Testing with the Ratio of Compression parameter shows that homogeneous characters have a smaller percentage value of 62.09% than heterogeneous characters 62.22%. The compression Ratio test for homogeneous characters has the same ratio as heterogeneous characters, namely 1.61. In space saving test, the percentage of homogeneous characters is 37.91% greater than heterogeneous characters at 37.78%. Testing based on Bitrate, homogeneous characters with heterogeneous characters has the same value of 4.97. Moreover, in The time comparison, homogeneous characters are faster than heterogeneous characters on a 1000-character test with a homogeneous time of 1.0907 ms and a heterogeneous time of 1.7975 ms.
Keywords:	AES 256, Cryptography, Compression, Levenstein Code
Source:	Journal of Theoretical and Applied Information Technology 30^th November 2023 -- Vol. 101. No. 22-- 2023
Full Text

Title:	AN OPTIMIZED HYBRID CUCKOO SEARCH BIO-INSPIRED DEEP BELIEF NETWORK FOR HEART DISEASE PREDICTION
Author:	M. SIVAKAMI, P. PRABHU
Abstract:	In today’s situation the heart diseases are one of the most common problems for many people who is working under pressure environments. These cause severe life losses for many of the aged persons. So nowadays many techniques used for the classification and the prediction of the heart diseases among the affected persons. In this novel the machine learning is used for computing the data’s directly from the ECD data and set the out of heart diseases predictions. Based on the analysis, about 55% of the men are affected by the heart diseases compared to the women. The idea behind the Deep Belief Network with suitable metaheuristic algorithm plays significant contribution to predict the heart disease. Neural convolution model with Adam Optimizer, Decision tree with grey wolf model, Support vector machine classifier using Bayesian Optimization techniques, cuckoo search bio inspired deep belief network is also used to predict heart disease with lesser accuracy. Tiny deviation accuracy in the prediction of heart disease causes to increase wrong treatment. This paper proposes the advanced technique of Cuckoo search bio inspired among deep belief network model with SVM classifier for the identification of the heart diseases. Hamming distance feature selection and Gaussian filter also deployed for preprocessing and data cleaning. The proposed optimized hybrid cuckoo search bio inspired Deep Belief Network (CS-DBN) improve accuracy to predict heart disease. Comparative analysis was made with existing Convolution Neural Model with Adam Optimizer (CNMAO), Decision tree with grey wolf Model (DT-GWN), Support vector Machine classifier with Bayesian Optimization algorithm (SVM-BOA), Convolution Neural network with Social Mimic Optimization (SVM-SMO). Compared to the previous statistics the proposed hybrid CS-DBN algorithm is one of the most efficient methods of the prediction and the classification of the heart diseases. Here about 99.5% accuracy of the disease’s prediction can be get using this advanced proposed method. For performing the analysis of the damaged ECG signal among all the ECG signal data which is collected from the data set is done. Then for this type of analysis the python platform is used for the disease’s predictions.
Keywords:	Support Vector Machine, Cuckoo Search Bio Inspired Model, Deep Belief Networks. Gaussian Filter
Source:	Journal of Theoretical and Applied Information Technology 30^th November 2023 -- Vol. 101. No. 22-- 2023
Full Text

Title:	ANALYSIS OF GAME THEORY IN MARKETING STRATEGIES OF TIKTOK AND INSTAGRAM
Author:	HENOCH JULI CHRISTANTO, STEPHEN APRIUS SUTRESNO, VALDI SEPTIAN SIMI, CHRISTINE DEWI, GUOWEI DAI
Abstract:	The rise of social media platforms like Instagram and Tiktok has had a profound impact on the business landscape, leading to a significant increase in e-commerce activities. This study focuses on examining the marketing strategies of Instagram and Tiktok using the 7P framework, which includes Product, Price, Promotion, Place, People, Physical evidence, and Process. Understanding these strategies is crucial for businesses aiming to enhance their online presence and drive sales.The study employed purposive sampling and collected data through questionnaires from individuals who frequently used Instagram and Tiktok for shopping. Validity was assessed using Karl Pearson's product-moment correlation, while reliability was evaluated using Cronbach's alpha. The data was then analyzed using Game Theory to identify optimal strategies.The findings highlight the significance of the "Place" element in the marketing strategies of Instagram and Tiktok. Both platforms prioritize user-friendly interfaces, secure and timely deliveries, and convenient product selection. The Game Theory analysis revealed a saddle point value of 2, indicating that the implemented strategies are effective in maximizing profits and minimizing losses. This research provides valuable insights into the marketing strategies of Instagram and Tiktok, demonstrating the relevance of the 7P framework and Game Theory in driving successful online businesses. By implementing these strategies, businesses can improve their competitiveness, attract a broader customer base, and thrive in the dynamic digital market. Understanding the effective marketing approaches employed by Instagram and Tiktok serves as a practical guide for aspiring entrepreneurs and existing businesses to navigate the realm of e-commerce and achieve sustainable growth.
Keywords:	Social Media, 7P Framework, Game Theory, Marketing Strategies
Source:	Journal of Theoretical and Applied Information Technology 30^th November 2023 -- Vol. 101. No. 22-- 2023
Full Text

Title:	DENSE INCEPTION ATTENTION NEURAL NETWORK FOR ALZHEIMER'S DISEASE CLASSIFICATION USING MAGNETIC RESONANCE IMAGES
Author:	K. RANGA SWAMY, S. SENTHILKUMAR
Abstract:	Alzheimers disease (AD) is an irreversible disease and all currently available treatments may only delay its progress. Nonetheless, the diagnosis of AD, especially in the early stages, is important for preventing it from affecting daily life. Many current deep neural network models are designed to improve performance by deepening the number of network layers and widths for AD diagnosis using MRI images. It is characterized by both cognitive and functional impairment. However, as AD has an unclear pathological cause, it can be hard to diagnose with confidence. To address the above problems, this paper proposes a Dense Inception Attention Neural Network (DIAN-Net), which is a combination of the Dense module, skin connection, triple attention block, and the Inception attention module. The dense network uses depth separable convolution to remove the redundant operations of conventional convolution, constructs feature-separated distillation blocks as the basic feature extraction blocks of the network to extract multi-level depth feature information, and combines triple attention (TA) to enhance the feature mapping capability of the network and reduce model parameters and computation. The inception attention module is used to improve the receptive field and thus capture global feature information. The discriminative network makes the model more attentive to detailed image features and stabilizes the training process by taking into account the effects of different parts of the image. The experimental results show that the algorithm in this paper has a smaller number of parameters and shorter training time, and outperforms other methods in terms of subjective visualization and objective evaluation metrics on multiple benchmark datasets.
Keywords:	Alzheimers Disease (AD), DCNN, Data Fusion, Dense Attention Network
Source:	Journal of Theoretical and Applied Information Technology 30^th November 2023 -- Vol. 101. No. 22-- 2023
Full Text

Title:	COMPREHENSIVE FUSION OF ADVANCED TECHNIQUES FOR PRECISE LUNG CANCER DETECTION IN HISTOPATHOLOGY IMAGES
Author:	DINESH NANNAPANENI, VEERA RAGHAVA SAI VARMA SAIKAM, RAJESH SIDDU, VAMSI MANOJ CHALLAPALLI, VENUBABU RACHAPUDI
Abstract:	Recent advances in deep learning have ushered in a new era in medical research, especially in the complex field of lung cancer identification in histopathology pictures. The innovative use of deep learning algorithms for locating lung cancer symptoms in histopathology material is thoroughly examined in this research. A promising path to improving the accuracy, efficacy, and thoroughness of identifying this potentially fatal condition emerges using artificial intelligence. As they carefully analyze large histopathology picture datasets and reveal the crucial characteristics closely connected to lung cancer pathology, the authors of this study set out on a rigorous journey to understand the potential relevance that deep literacy models may offer. The desired result is a significant improvement in the precision and promptness of diagnostic assessments, which would significantly improve patient care procedures. This work aims to improve the understanding of the scientific community by traversing the complex abstractions inside histopathology-based lung cancer image analysis via the perspective of deep literacy. The research's forward momentum extends to shedding light on a game-changing approach for more advances in the field. It is hoped that by pursuing this research, early detection techniques and cutting-edge treatment approaches would develop, especially for those dealing with lung cancer. This research is a significant step forward in understanding lung cancer through images of tissue samples. The new methods explored here have the potential to greatly improve how we diagnose this disease, leading to better outcomes for patients. A new era of improved patient outcomes and top-notch healthcare may be started by reshaping the medical research environment thanks to the synergy between deep literacy and histopathology.
Keywords:	Image detection, Histopathology images, Lung cancer, Imaging, Lung cancer, Predictions, Clinical outcomes.
Source:	Journal of Theoretical and Applied Information Technology 30^th November 2023 -- Vol. 101. No. 22-- 2023
Full Text

Title:	SMART AND SECURED SOFTWARE DEFINED NETWORKS WITH MCELIECE CRYPTOSYSTEM FOR INTERNET OF THINGS USING BLOCK CHAIN
Author:	J.V.N RAGHAVA DEEPTHI, AJOY KUMAR KHAN, TAPODHIR ACHARJEE
Abstract:	The term Internet of Things (IoT) is used to describe everyday objects that can connect to and share data with other systems and devices over the Internet or other forms of communication with built-in sensors, processors, software, and other technologies. The primary objective is to plan and build an SDN-based IoT ecosystem with a secure communication infrastructure. The McEliece cryptosystem is used to further encrypt the block chain mechanism, and this implementation phase also includes intelligence into the data transfer mechanism. When data consisting of many processes related to controlling, routing, managing logic, etc. is sent to the SDN block, the complexity of the underlying mechanism is reduced. In a software-defined network, the data is stored in a block chain, encrypted with McEliece encryption, and then authenticated across the whole control, data, and application layers. The control plane is also used for operations and forwarding, allowing for better decision making. This study explains how SDN may be used to build a smart computational model with a wider range of features, all while keeping the underlying communication system secure. This article proposes a solution that uses the blockchain and the McEliece cryptosystem in an optimal layout to provide unprecedented levels of safety.
Keywords:	Block chain, IoT, SDN, encryption, security, delay, node, cryptosystem, data transmission, and McEliece.
Source:	Journal of Theoretical and Applied Information Technology 30^th November 2023 -- Vol. 101. No. 22-- 2023
Full Text

Title:	HOW CAN TRANSFORMER-BASED BIDIRECTIONAL ENCODERS ENHANCE THE CLASSIFICATION OF STUDENT PUBLICATIONS ON SOCIAL MEDIA?
Author:	OMAR ZAHOUR , EL HABIB BENLAHMAR , MEHDI EL AMINE
Abstract:	Text categorization, especially concerning student submissions, constitutes a fundamental undertaking in the realm of natural language processing (NLP). In recent times, bidirectional encoder representations from transformers, commonly known as BERT, a renowned model developed by Google, has been at the forefront in yielding groundbreaking outcomes across a plethora of NLP tasks. In the contemporary era, where text forms the bulk of the data available globally, the significance of automated NLP techniques cannot be understated, emerging not just as a valuable tool but a critical necessity in the sphere of artificial intelligence. While models like BERT have made notable strides, offering unprecedented results in comparison to earlier methods, they sometimes overlook the nuanced local information embedded within the textual data, such as inter-sentential relations. In this study, we delve into the intricacies of leveraging BERT for multi-class text classification, focusing specifically on categorizing student articles pertinent to the educational and vocational guidance sector, in alignment with Holland's RIASEC theory. Our constructed model is adept at determining the precise category of the input publication, operating within a framework that recognizes six distinct classes encompassing our dataset. Conducted through comprehensive experiments utilizing Python, the proposed model exhibits a promising performance benchmark, standing as a potent contender in the ever-evolving domain of text classification and NLP.
Keywords:	Text Categorization; Bidirectional Encoder Representations from Transformers (BERT); Natural Language Processing (NLP); Multi-class Classification; Educational and Vocational Guidance; RIASEC Typology of Holland
Source:	Journal of Theoretical and Applied Information Technology 30^th November 2023 -- Vol. 101. No. 22-- 2023
Full Text

Title:	ABC ALGORITHM AS AN ENHANCEMENT FOR MQO PROCESS IN BIG DATA
Author:	MANAL A. ABDEL-FATTAH , SAYED ABDELGABER , S. A. NASR , WAEL MOHAMED
Abstract:	Multi-query optimization is the task of generating an execution plan for a collection of multiple queries. In recent years, big data querying has become an important field because it provides better data understanding and valuable insight. This paper studies the ability to enhance the process of multi-query using one of the swarm algorithms. Join operation is the most time-consuming operation, the study focuses on join operation and illustrates the effect of the join execution order on the time. Many techniques can be used to decide the optimal order to execute a set of join operations, swarm algorithms are proposed in this research to scan all possible solutions and choose the optimal one. The paper provides a model for the process, examines it on the big data set, and compares it with previous work. The experiment results that applying artificial bee colony algorithm on multi-query optimization enhances the time of execution.
Keywords:	Multi-Query Optimization (MQO), Query Execution Plan (QEP), Artificial Bee Colony (ABC)
Source:	Journal of Theoretical and Applied Information Technology 30^th November 2023 -- Vol. 101. No. 22-- 2023
Full Text

Title:	A GAMIFIED E-LEARNING MODEL BASED ON THE ONE SIZE FITS ALL MODEL AND THE STATIC AND DYNAMIC ADAPTATION MODEL
Author:	ZECRI EZZOUBAIR , OUZZIF MOHAMMED , EL HADDIOUI ISMAIL
Abstract:	In order to create a gamified e-learning system, several researchers have proposed models. some are based on the use of the same game elements for all users / learners "one size fits all", others have opted for the modification of these elements depending on the learner, we are talking about adaptive gamification or personalized gamification, this second type of model can be divided into two categories which are the static adaptation models in which the gamified e-learning system offers certain elements according to learners information or their learning style. The second category is the dynamic adaptation models which propose game elements depending on the behavior of the student during his learning process. In this paper we propose a new model in which our gamified learning system will display the game elements using the three categories explained above according to the type of game element, if they are easily noticed by the learner (intrinsic) or not ( extrinsic).The objective is to help designers of gamified LMS and teachers to choose the right games element to display on the course in order to better maintain the learner motivation.
Keywords:	Adaptive Gamification, E-Learning, Instructional Design, Adaptation Model
Source:	Journal of Theoretical and Applied Information Technology 30^th November 2023 -- Vol. 101. No. 22-- 2023
Full Text

Title:	COMPARISON OF NAÏVE BAYES AND LIWC FOR SENTIMENT ANALYSIS OF GOJEK (GOTO FINANCIAL) USER SATISFACTION
Author:	EZRA PERANGINANGIN , EMIL R. KABURUAN , YEFTA ANDREA
Abstract:	Customer satisfaction impacts customer retention. Thus, it is important for the service provider’s business development including for GoTo Financial which has a very large number of subscribers. Current technological advances have provided customers with many channels to submit complaints, criticisms, compliments, and all kinds of opinions to service providers. Not only through customer service agents, but social media like Twitter are also frequently used by customers to reach out to the provider. Google Play Store also has an Application Review feature that customers often use to express their experience using the application. This study analyzes the customer satisfaction level of GoTo Financial. Two methods namely Naive Bayes and Lexicon (LIWC) are employed to obtain three sentiment classes i.e., negative, neutral, and positive, with data obtained from Twitter and Google Play Store based on the most frequently complained topic through the company's customer service. With one topic and a limit of 1000 data, the results show that this company is proven to have many daily active customers. The results also suggest that Lexicon’s accuracy is low at 7.43% due to the large number of false negatives. In contrast, the Naive Bayes multinomial consistently shows high accuracy at 87.76% over Twitter data and 71.25% over Google Play Store data, making it a better method than Lexicon.
Keywords:	GoTo Financial; Naive Bayes; LIWC; Sentiment Analysis, User Satisfaction
Source:	Journal of Theoretical and Applied Information Technology 30^th November 2023 -- Vol. 101. No. 22-- 2023
Full Text

Title:	AN INTEGRATED ITERATIVE FEATURE SELECTION AND NAIVE BASED ORIENTED SUPERVISED MACHINE LEARNING TECHNIQUE TO DETECT SQL INJECTION ATTACKS IN IOTS BASED COMMUNICATION
Author:	ASIFIQBAL SIRMULLA , PRABHAKAR M
Abstract:	IoT is a type of networking model which is composed of several wireless and wired networks that are interconnected through internet-based channels. These devices are widely adopted in various types of applications such as home automation, industries and academic purpose. However, this type of easily accessible connectivity has unlocked several challenging issues where maintaining secure and reliable connectivity is a crucial task. In these attacks, injecting the false queries is considered as one of the most challenging security attacks announced by the OWASP and SQL injection is the common type of injection attack. To deal with this issue, machine learning is considered as promising technique which learns the pattern of historical data and detect the attack by processing the suspicious queries. In this work, we present a supervised learning-based machine learning algorithm which is based on the Naïve Bayes classifier. We present a threefold strategy in proposed scheme, which contains a feature modelling phase where a mutual information based feature dependency model is utilized. Later, an iterative feature selection model is also presented to select features iteratively until the feature selection criteria is obtained. Finally, a naïve Bayes classifier model is presented to classify the selected attributes. The performance of proposed approach achieves the average performance as 0.96, 0.958, 0.97, and 0.979 in terms of Avg. Precision, Avg. Recall, F1-Score, and Accuracy.
Keywords:	SQL Injection Attacks, Security, Naïve Bayes, IoT, Kaggle, SQLIA
Source:	Journal of Theoretical and Applied Information Technology 30^th November 2023 -- Vol. 101. No. 22-- 2023
Full Text

Title:	AN IMPROVED PROPORTIONAL RECONFIGURABLE INTEGRATED DECODING WITH BELIEF PROPAGATION (PRDBP) FOR 5G COMMUNICATIONS
Author:	PECHETTI GIRISH , BERNATIN T
Abstract:	Recently, Fifth Generation (5G) communication network uses the Low-Density Parity Check (LPDC) codes due to minimized latency in the radio channel. The conventional LPDC uses the standard decoders that use an iterative message-passing (IMP) algorithm with iterative decoding based on belief propagation (IDBP) or the sum-product algorithm (SPA). Those iterative message algorithms are operated in a loopy graph with the maximum-likelihood estimation which leads to higher computation complexity for the dynamic update. To overcome the issues proportional reconfigurable decoding model is utilized in place of an iterative algorithm to perform low-density parity check (LDPC) decoding. Additionally, the 5G network demands for higher throughput, compatibility and minimal complexity. This paper proposed an Integrated Proportional Reconfigurable Decoding with Belief propagation (PRDBP) to increase the throughput in the 5G wireless communication channel. The proposed PRDBP model focused on linear error-correcting codes in 5G channels. The model computes the enumerator’s code word length for the finite length of code ensembles. The PRDBP model considers the block-length computation for the asymptotic cases to compute the minimal sensing distance. The code ensemble in the PRDBP model characterized the trapping set, stopping set, and properties of the code word in minimal floors. The proposed PRDBP model integrates the sum products and belief propagation algorithm for the modification in node equations. Based on certain class interior point estimation method linear problem and nonlinear convex optimization problems are computed. Based on computation affine scaling, reduction method and path-following methods are implemented.
Keywords:	Low Density Parity Check, Iterative Message-Passing, Proportional Reconfigurable Decoding, Fast Fourier Transformation, Progressive Edge Growth, Sum Product Algorithm.
Source:	Journal of Theoretical and Applied Information Technology 30^th November 2023 -- Vol. 101. No. 22-- 2023
Full Text

Title:	HIGHLY RELIABLE ULTRA-LOW POWER AND LATENCY OPTIMISED FIN-FET BASED 9T SRAM
Author:	SHYAM.K , DR.V. VIJAYAKUMAR
Abstract:	The limitations of CMOS at smaller scales have increased the demand for alternate nano-devices. There are a number of suggested devices, including FinFET, TFET, and CNTFET. FinFET stands out as a promising device that has the potential to replace CMOS due to its low leakage in the nano meter range. Today data storage very much important in daily life. So, for this we required sophisticated electronic devices for storage purpose. So, we have to design basic SRAM Cell because of SRAM has less static Power compared to DRAM. To strike a better balance between the improved transistor circuit design and the increased storage capacity required by modern electronics, a new technology known as FinFET has been created. The loss of gate control over the channel causes many problems for CMOS devices, including higher production costs, higher ON current, short channel effects (SCEs), lower reliability and yield, higher leakage currents, and so on. This idea proposes a unique FinFET based 9T SRAM cell that uses a single ended bit-line design for near-threshold read/write operations. This eliminates the need for a boosted power supply and writes aid circuitry. As well as improving write-ability, write power, and write time, turning off both M4 and M5 transistors in write mode also removes write and read limits on the size of the semiconductor device. By using a transmission gate with a low threshold voltage (Vt) as the access transistor, both write-ability and write-time were dramatically enhanced.
Keywords:	System-on-chip, read static noise margin, write access time, leakage power, Carbon Nano Tubes, 9TSRAM
Source:	Journal of Theoretical and Applied Information Technology 30^th November 2023 -- Vol. 101. No. 22-- 2023
Full Text

Title:	MULTIMODAL LEARNING CONVERSATIONAL DIALOGUE SYSTEM: METHODS AND OBSTACLES
Author:	MUHAMMAD FIKRI HASANI , GALIH DEA PRATAMA, ERNA FRANSISCA ANGELA SIHOTANG, FRANZ ADETA JUNIOR
Abstract:	Humans converse with each other as one of the means to interact socially. But conversation not only served as media of communication, but also opened several occupancies like moderators, customer services, master-of-ceremony, or teachers. Dialogue system is a computer program that supports spoken, text-based, or multimodal conversational interactions with humans with many implementations recently and only work in single modality, such as text. However, human understanding is not limited into single data domain, but it needs the collective data domain information to understand the whole surroundings which is called as multimodality in the field of computer science and implemented further through the concept of artificial intelligence called multimodal learning. Multimodal learning has been subjected in research since years ago to increase artificial intelligence model result, such as enhancement of speech recognition through mouth image and facial expression recognition based on facial and landmark textures. This paper will provide reference in integration of multimodal learning in dialogue system, which will be useful to negate obstacles present in future research.
Keywords:	Multimodal learning, Dialogue system, Systematic literature review, Chatbot, Deep learning
Source:	Journal of Theoretical and Applied Information Technology 30^th November 2023 -- Vol. 101. No. 22-- 2023
Full Text

Title:	UNDERSTANDING FACTORS AFFECTING USER BEHAVIOR INTENTION ON LIVE STREAMING SERVICES ON MARKETPLACE IN INDONESIA
Author:	RICKY REYNALDY, ASTARI RETNOWARDHANI
Abstract:	Along with the times and the development of technology, users can now use the internet in various ways, such as a recent phenomenon, namely the rise of live streaming services on various applications such as marketplace applications, especially in Indonesia. Seeing the high enthusiasm and potential for sales and purchases on live streaming services, this study was made to understanding the user's desire to use it in the future, the desire to use it frequently, the desire to use it in daily life, and the desire to use it regularly play an important role in the continued use of this new technology, and to find out what factors that affecting behavior intention of users when they use live stream services on several marketplace applications in Indonesia including Tokopedia, Shopee, and Lazada. This study uses a questionnaire distribution method to collect data and uses the Technology Acceptance Model (TAM) as a research model. The results shows that perceived interaction and perceived usefulness have a significant effect on behavior intention, while perceived ease of use does not have a significant effect on behavior intention.The results of this study will generate a new understanding of the behavior intention of users when using the live stream marketplace service and allow for continuous research from the related scope.
Keywords:	Live Streaming, Marketplace, Behavior Intention, Technology Acceptance Model (TAM), Perceived Ease Of Use, Perceived Usefulness.
Source:	Journal of Theoretical and Applied Information Technology 30^th November 2023 -- Vol. 101. No. 22-- 2023
Full Text

Title:	ONLINE SOCIAL INTERACTION COMMUNITY MODEL FOR UNMET NEEDS OF OLDER PEOPLE
Author:	FARHAT EMBARAK, NOR AZMAN ISMAIL, MASITAH GHAZALI, MURNI MAHMUD, NUR ZURAIFAH SYAZRAH OTHMAN, LAYLA HASAN, CHE SOH SAID, PANG YEE YONG, ALHUSEEN OMAR ALSAYED
Abstract:	Modern medicine and technology have enhanced the average life expectancy of a person, but this has not improved the quality of life for those who are getting older. Elderly persons struggle to take care of their fundamental requirements, and family members are unable to give them the attention they require. An online community based new method of communication and interaction for the elderly could be a solution to address the situation, however, there are difficulties observed with the elderly persons throughout the design stage. In order to address the unmet everyday needs of older adults, this study investigated the design methodology and design process for an online social interaction community (OnSocialCom). A conceptual model was proposed specifically for senior people who live alone and was based on social connectedness, social support models, structural Camberwell evaluation of need for the elderly theories, and 4W (What-Where-When-Who) models. Its application in the actual setting was investigated through empirical investigations involving users and experts. The developed model contains five modules; online social interaction, profile management, unmet needs interaction, unmet needs plan, and recommendation. Additionally, evaluation was performed during the design process by experts for review analysis to clarify the needs of the users and assist in optimising the functionality of the conceptual model.
Keywords:	Elderly People, Social Interaction, Unmet Needs, Users, Social Connectedness, Recommendation.
Source:	Journal of Theoretical and Applied Information Technology 30^th November 2023 -- Vol. 101. No. 22-- 2023
Full Text

Title:	GREEN TRAFFIC ENGINEERING : SOLUTION FOR REDUCING ENERGY CONSUMPTION IN SDN NETWORK WITH SEGMENT ROUTING
Author:	FARAH LAMHARRAS , NAJIB ELKAMOUN ,OUIDAD LABOUIDYA ,HIND SOUNNI
Abstract:	Thanks to its flexibility of handling and processing due to the separation of control and data planes, The new Software Defined Network (SDN) paradigm has attracted a lot of attention from researchers. On the other hand, energy consumption in ICTs has become an important area of research, given the high energy dissipation due to the manufacture of high-tech equipment and the direct use of electricity. The aim of this study is to improve the energy efficiency of networks by switching off a subset of the links, using an SDN (Software Defined Network) approach.We dynamically adapt the number of network links that are switched on according to the network load.Our solution is based on the Segment Routing protocol, which we will demonstrate that is better than MPLS.The algorithms were implemented and evaluated using the OMNET++ simulator.Experimental results show that the number of links on can be reduced by a very interesting percentage while maintaining a high quality of service and network stability.the proposed solution presents a promising results to address the environmental and energy challenges faced by modern networks.
Keywords:	SDN, Green,Traffic Engineering,Energy saved,Segment Routing
Source:	Journal of Theoretical and Applied Information Technology 30^th November 2023 -- Vol. 101. No. 22-- 2023
Full Text

Title:	TBVA-GAN: A CLASS ACTIVATION MAPPING-GUIDED TUBERCULOSIS VISUAL ATTRIBUTION GENERATIVE ADVERSARIAL NETWORK
Author:	DING ZEYU, RAZALI YAAKOB, AZREEN AZMAN, SITI NURULAIN MOHD RUM, NORFADHLINA ZAKARIA, AZREE SHAHRIL AHMAD NAZRI
Abstract:	Visual attribution (VA) methods play a crucial role in tuberculosis (TB) research by providing valuable insights into disease patterns and aiding in diagnostic interpretation. The advent of generative adversarial network (GAN)-based VA methods has gained significant attention from researchers due to their ability to generate fine-grained feature maps that accurately reflect the location of lesions. These methods localize lesions by converting chest X-ray (CXR) images containing lesions into normal CXR images and analyzing the differences between the two. However, current methods only perform surface-level transformations, neglecting the vital information of whether lesions are present. Moreover, the transformation process assigns equal weights to the entire image, without specifically prioritizing the regions with a higher probability of lesions occurrence. In this study, a novel framework is proposed, namely the class activation mapping-guided tuberculosis visual attribution generative adversarial network (TBVA-GAN). This innovative model leverages the informative regions derived from class activation mapping to effectively guide the GAN in prioritizing the transformation of these crucial areas. Moreover, to guarantee the precision of TB localization, an auxiliary TB detection model is incorporated, ensuring that the converted CXR images are devoid of TB pathology. By employing this additional verification step, the accuracy of TB localization is significantly enhanced. The proposed TBVA-GAN in this study achieves promising VA results on the TBX11K dataset, surpassing existing GAN-based TB VA models.
Keywords:	Visual attribution, Tuberculosis, Deep learning, GAN
Source:	Journal of Theoretical and Applied Information Technology 30^th November 2023 -- Vol. 101. No. 22-- 2023
Full Text

Title:	REAL-TIME CLOUD-BASED AUTOMATION FOR CYBER THREATS DETECTION AND MITIGATION WITH MACHINE LEARNING MODELS
Author:	YOUNES WADIAI, YOUSEF EL MOURABIT, MOHAMMED BASLAM3, BOUJEMAA NASSIRI, YOUSSEF ELHABOUZ
Abstract:	Using Cloud Computing in the artificial intelligence field is a paramount tool applied in technology solutions nowadays to build strong machine learning models. Most Cloud solutions available in the technology market offer gigantic storage spaces with high computing performances that are easily accessible online via terminals where local machines cannot compete with the performance of these cloud resources. Machine learning algorithms are becoming increasingly popular in the field of cloud security, as they provide powerful tools for detecting and mitigating a wide range of cyber threats. However, the full exploitation of web service portals offered in these cloud platforms is still limited and it is applied mostly to gain access or retrieve data from databases stored in servers. Using cloud services is beneficial and can allow the automation of the process of building trained models in real-time. In this article, we propose the usage of Microsoft Azure Machine Learning Studio Web Service tool to train manly the Multi Class Neural Network model and other algorithms using the CSE-CIC-IDS2018 dataset. While the experiments are performed using the attributes from the predefined dataset, the feeding process is conducted by building, deploying, and running simulated real-time collected attributes from the IDS/IPS system.
Keywords:	Web Services, Cloud Computing, Machine Learning, Deep Neural Network, CSE-CIC-IDS2018 Dataset, Real-time Filtering, Synchronous Learning
Source:	Journal of Theoretical and Applied Information Technology 30^th November 2023 -- Vol. 101. No. 22-- 2023
Full Text

Title:	REPEATED NODE BEHAVIOUR ANALYSIS WITH NODE TRANSMISSION PATTERN ANALYSIS BASED BEHAVIORAL INDEX FOR FALSE ALARM DETECTION IN WIRELESS SENSOR NETWORKS
Author:	D. MURLI KRISHNA REDDY, DR. R. SATHYA, DR.V.V.A.S. LAKSHMI
Abstract:	Multiple types of sensor failures and inaccurate readings can compromise the integrity of a Wireless Sensor Network (WSN). The inability to quickly and accurately respond to emergencies is a major flaw in many WSN applications. In this research, a unique method for identifying sensor abnormality through the examination of physiological data gathered is proposed. Wireless sensor networks have a number of issues, one of the most significant being security. This research focuses mainly on the characteristics of a layered sensor node and its application in an intrusion detection system due to energy and processing constraints. The method's purpose is to accurately discern between false and true intrusion alarms. It does a comparison between the calculated sensor value and the current reading. The sensor reading is compared to a moving threshold value that indicates whether or not the reading is abnormal. This paper examines strategies for dynamically and efficiently decreasing the possibility of false alarms while increasing the likelihood that no target would go unnoticed. The proposed method adjusts the false positive rate threshold up or down when the false positive rate changes. The likelihood of identifying false alarms improves as a result. The findings recommend pooling data from multiple sensors to produce a comprehensive analysis of the target while keeping intrusions to a minimum. The nodes in the WSN are vulnerable to several sorts of intrusions. The node behaviour analysis is performed to identify the node actions frequently in the WSN. The nodes that are causing false alarms need to be identified for improving the network performance levels. This research presents a Repeated Node Behaviour Analysis with Node Transmission Pattern Analysis based Behavioral Index (RNBA-NTPA-BI)model is proposed for reduction of false alarms in the network. The proposed model is compared with the traditional model by considering the evaluation metrics like node behaviour analysis accuracy levels, false alarms detection accuracy levels. The proposed model exhibits better performance when compared to the traditional models.
Keywords:	Wireless Sensor Networks, Intrusion Detection, False Alarms, Node Behaviour Analysis, Behavioral Index, Network Performance.
Source:	Journal of Theoretical and Applied Information Technology 30^th November 2023 -- Vol. 101. No. 22-- 2023
Full Text

Title:	UNVEILING THE RESEARCH IMPACT: A VISUALIZATION STUDY OF CHATGPT'S INFLUENCE ON THE SCIENTIFIC LANDSCAPE
Author:	JEENA JOSEPH , JOBIN JOSE , ANAT SUMAN JOSE , GILU G ETTANIYIL , JOSHY JOHN , PRINCY D NELLANAT
Abstract:	Numerous industries, including research, have undergone a radical transformation because to the creation of cutting-edge NLP models. These models are very interested in ChatGPT due of its remarkable language production capabilities, which OpenAI developed. To assess the effect of ChatGPT on the research landscape, this study performs a bibliometric analysis utilizing Citespace, VOSviewer, and Biblioshiny. The Scopus database has been thoroughly explored for relevant publications published after ChatGPT's launch. Bibliometric methods like co-citation, authorship, and keyword analysis were used to analyze the retrieved dataset. The analysis's findings reveal a startling rise in papers that mention ChatGPT, demonstrating the tool's rising stature in the scientific world. Some major research trends have been discovered by analyzing the most popular keywords for ChatGPT research. The results demonstrate the interdisciplinary nature of ChatGPT research and its incorporation into different academic subjects. The findings give academics and industry professionals a thorough picture of existing ChatGPT research, making it easier to identify areas that can benefit from more research and encouraging interdisciplinary collaboration.
Keywords:	ChatGPT, Bibliometric Analysis, Natural Language Processing, Artificial Intelligence, CiteSpace, VOSviewer, Biblioshiny.
Source:	Journal of Theoretical and Applied Information Technology 30^th November 2023 -- Vol. 101. No. 22-- 2023
Full Text

Title:	PERFORMANCE EVALUATION OF DTN ROUTING PROTOCOLS ON MAP-BASED SOCIAL MOBILITY MODELS FOR DTN NETWORKS
Author:	EL MASTAPHA SAMMOU
Abstract:	Delay-tolerant networks (DTNs) are mobile networks that experience frequent and persistent partitions due to high node mobility and sparse distribution of nodes. As well as, the unpredictable and random topology changes and frequent node disconnections encountered in DTNs pose several routing challenges and several challenges to the design of effective DTN routing-protocols. For the majority of DTN routing-protocols, nodes rely on their mobility to forward messages to their destinations. Therefore, it is important to understand the impact of commonly used social mobility models on the performance of DTN routing-protocols already designed for DTN networks. The main objective of this article is the evaluation of the performance of routing-protocols DTN, Epidemic, Prophet, MaxProp and Spray-and-Wait on the three map-based social mobility models, MBM (Map Based Movement), SPMBM (Shortest Path Map Based Movement) and MRM (Map Route Movement), taking into account four performance metrics: Delivery Rate, Average latency, Overhead Ratio and Average hop count. The performance evaluation and the production of realistic traces of the mobility of the nodes is done using the ONE (Opportunistic Network Environment) simulator. The simulation results obtained according to the density of the nodes show that the models of social mobility have a significant effect on the routing process and the performance of DTN routing-protocols.
Keywords:	Delay Tolerant Networks (DTN); Routing Protocol Performance; Epidemic; Prophet; MaxProp; Spray-and-Wait; Social Mobility Models; MBM; SPMBM; MRM; ONE Simulator.
Source:	Journal of Theoretical and Applied Information Technology 30^th November 2023 -- Vol. 101. No. 22-- 2023
Full Text

Title:	THE SUPERIORITY OF YOLOV4 MODEL FOR ENHANCING PALM OIL FRUIT DETECTION
Author:	VINCENT FANDITAMA WIJAYA, PATRICIA CITRANEGARA KUSUMA , BENFANO SOEWITO
Abstract:	This paper addresses challenges by investigating the integration of color and texture information to enhance object detection. We conducted a comparative analysis of various YOLOv4 models, including YOLOv4, YOLOv4 Tiny, YOLOv4 Tiny_3l, and YOLOv4_csp. Our study primarily employs the YOLOv4 model for image detection and focuses on experiments using the oil palm fruit bunch dataset. This dataset is driven by the potential to leverage texture and color analysis to assess the maturity level of oil palm fruits. Our research objectives center on evaluating how color and texture impact object recognition and exploring the capabilities of YOLOv4 models in this regard. The dataset consists of 4156 images categorized into 6 classes: Overripe, Ripe, Raw, Underripe, Abnormal, and Empty. Our experimental findings reveal that the YOLOv4 model excels in accurately identifying the color and texture attributes of oil palm fruit bunches. Notable performance metrics include an average IoU of 89.95%, mAP at 50 of 99.85%, recall of 0.996, F1-scores of 0.987, and precision of 0.979. These results underscore the superior capabilities of the YOLOv4 model in object recognition within this context. The significance of our results lies in advancing our understanding of how the integration of color and texture information can augment object detection. Furthermore, our findings highlight the efficacy of YOLOv4 models in this specific application, emphasizing their potential for broader applications in the field of object detection.
Keywords:	Object detection, YOLOv4, Color, Texture, Palm Fruit
Source:	Journal of Theoretical and Applied Information Technology 30^th November 2023 -- Vol. 101. No. 22-- 2023
Full Text

Title:	HBFZKP: A DUAL PRIVACY PROTECTION MODEL LOCATION BASED SERVICE SYSTEMS
Author:	KHALID ALSUBHI
Abstract:	Location-based services (LBS) have become pervasive, reaching all smart devices equipped with GPS, delivering substantial value to consumers. Despite their popularity, LBS has its shortcomings, particularly concerning the necessity for users to disclose their location data to fully benefit from the service, potentially compromising their privacy and security. As a result, numerous techniques have been proposed in the literature to offer an optimal solution for preserving privacy in LBS queries. This study will initially delve into three established approaches commonly employed for safeguarding privacy in location-based services: Zero Knowledge Proof, Oblivious Transfer, and the Bloom filter. Each of these methods aims to minimize the disclosure of information while simultaneously establishing an automated performance metric. Among the three methods mentioned, Bloom filters arguably exhibit the most efficient runtime performance. Nonetheless, Bloom filters exhibit two drawbacks: (a) they leak a maximum of one bit of information per query, and (b) the hash functions (Hk) require meticulous design and security analysis to ensure they are orthogonal and independent. This signifies that if one H is compromised (secure), nothing can be deduced about any other H, preserving the integrity of the remaining hash functions. In response to these challenges, we propose an innovative two-phase privacy-preserving framework for LBS, hybrid of Bloom filter and Zero Knowledge Proof (HBfZkp). While all three of these methods have demonstrated their efficacy in safeguarding a user's private data, our proposed two-phase privacy approach is poised to elevate privacy protection further. It is designed to shield users from both internal and external threats by capitalizing on the inherent strengths of Zero Knowledge Proof and Oblivious Transfer in safeguarding against these distinct types of attacks, respectively.
Keywords:	LBS, Privacy, Bloom Filter, Zero Knowledge.
Source:	Journal of Theoretical and Applied Information Technology 30^th November 2023 -- Vol. 101. No. 22-- 2023
Full Text

Title:	BUILDING A RETRIEVAL-AUGMENTED GENERATION SYSTEM FOR ENHANCED STUDENT LEARNING: CASE STUDY AT PRIVATE UNIVERSITY
Author:	M BAGASKORO TRIWICAKSANA S , TANTY OKTAVIA
Abstract:	This research conducted at Private University investigates the development and implementation of a Retrieval-Augmented Generation (RAG) system for enhanced student learning. The RAG system is a blend of retrieval-based and generative models using ChatGPT, aiming to address the challenges students face in accessing and understanding digital literature, mainly due to language barriers and passive reading methods. The RAG prototype was successfully created and assessed through black box testing and usability testing among students at Private University. Findings show that the RAG system significantly enhances interactive learning by providing contextually relevant answers. The system is highly functional and easy to use and can answer questions quickly and accurately. These results underscore the potential of the RAG system in transforming the educational process by offering an efficient and interactive learning and literature comprehension experience. This research highlights the need for further refinements in such systems, emphasizing their importance in educational settings.
Keywords:	Retrieval-Augmented Generation (RAG), Retrieval-based, Generative models, Student Learning
Source:	Journal of Theoretical and Applied Information Technology 30^th November 2023 -- Vol. 101. No. 22-- 2023
Full Text

Title:	VARIABLE LENGTH PACKET CIPHER USING CATALAN SEQUENCE
Author:	V. UMA KARUNA DEVI KAKARLA, CH. SUNEETHA
Abstract:	In recent times, the world is transforming to digital communication from than physical communication. Secure data transfer has become essential and challenging task all over the world. Cryptography is the science of secure communication of sensitive data via public channel. Encryption algorithms use mathematical techniques to create confusion and non-comprehensible to unintended persons. Applied number theory and cryptography have inextricable attachment. Many tools of elementary and applied number theory have vast applications in cryptography. The present paper aims at designing a variable length packet cipher using Catalan number sequence. Sequence of Catalan numbers forms variable size matrices when arranged in a special pattern. The interesting fact is that all these matrices are symmetric having determinant one. These matrices are used in the present algorithm for encryption and decryption.
Keywords:	Catalan Number Sequence, Encryption, Decryption, Matrices.
Source:	Journal of Theoretical and Applied Information Technology 30^th November 2023 -- Vol. 101. No. 22-- 2023
Full Text

Title:	A DETAILED REVIEW ON MULTI-MODALITY BASED EMOTION DETECTION WITH PHYSIOLOGICAL SIGNALS USING DEEP LEARNING
Author:	T L DEEPIKA ROY, D NAGA MALLESWARI
Abstract:	In several fields, such as healthcare, human-computer interaction, and affective computing, emotion recognition has attracted a lot of attention. This review paper investigates the new area of multi-modality emotion detection, which uses deep learning methods and physiological signals to improve the precision and resilience of emotion identification. We give a summary of the basic ideas, approaches, difficulties, and possible uses in this multidisciplinary field. Additionally, we examine the difficulties and constraints encountered in multi-modal emotion detection, encompassing the gathering of data, the extraction of features, and the interpretability of deep learning models. We highlight trends and new directions in the field in this study, including real-time applications, cross-modal emotion recognition, and transfer learning. A summary of the state of the art is provided in the paper's conclusion, highlighting the potential significance and real-world applications of multi-modality-based emotion detection with physiological signals through deep learning. Future directions for research in this exciting and rapidly developing field are also outlined.
Keywords:	EEG, ECG, FACS, HRV, AU…
Source:	Journal of Theoretical and Applied Information Technology 30^th November 2023 -- Vol. 101. No. 22-- 2023
Full Text

Title:	DEEP YOLOv8-BASED HANDBALL DETECTION SYSTEM WITH TRANSFER LEARNING APPROACH
Author:	R.J.POOVARAGHAN, P. PRABHAVATHY
Abstract:	Utilizing advanced computer vision techniques such as deep learning and object tracking algorithms, AI-powered active player detection in handball videos offers the capability to automatically track players' movements within high-speed matches. This innovation not only enriches coaching insights into player performance and team dynamics but also elevates viewer engagement through real-time analysis and augmented reality enhancements. In the context of practice-based handball videos, where multiple players frequently appear together, not all participants engage in the specific exercise or adopt the recommended handball techniques. This study explores the novel approach of employing the CNN based YOLOv8 pre-trained model in conjunction with transfer learning techniques for enhanced handball recognition. The YOLOv8 architecture's advanced capabilities are harnessed to address existing gaps in player tracking, ball trajectory prediction, and complex player interactions. Through transfer learning, the model is fine-tuned using handball-specific data, enabling adaptation and specialization in identifying players, the ball, and key elements. The method leverages YOLOv8's real-time processing and multi-scale analysis to improve accuracy in dynamic game scenarios, overcoming challenges like occlusion and rapid motion. By integrating the YOLOv8 pre-trained model with transfer learning, this approach showcases a promising advancement in achieving comprehensive and efficient handball recognition, significantly enhancing insights into player dynamics, ball movement, and overall gameplay. The fusion of YOLOv8 with transfer learning involves leveraging YOLOv8's pre-trained features for extracting object characteristics, followed by fine-tuning the model on handball-specific data to enhance its ability to recognize players, the ball, and other essential elements in the context of handball recognition. We systematically evaluated the proposed approach using a custom dataset of 751 handball scene videos captured during training sessions for young cadets and handball schools for both girls and boys [22]. Testing encompassed nearly 60,000 frames and incorporated metrics such as sensitivity, specificity, and accuracy. The results demonstrated that our method surpassed state-of-the-art techniques, showcasing heightened accuracy. Notably, the proposed method exhibited enhanced efficiency, achieving a sensitivity 92.18%, specificity 91.13%, accuracy 93.57% and F-score 94.33% respectively.
Keywords:	Handball Recognition, Deep Learning, YOLOv8, Transfer Learning, Computer Vision
Source:	Journal of Theoretical and Applied Information Technology 30^th November 2023 -- Vol. 101. No. 22-- 2023
Full Text

Title:	TOWARDS A HYBRID APPROACH TO REVERSE ENGINEER BEHAVIORAL UML DIAGRAMS FROM SOURCE CODE
Author:	HAMZA ABDELMALEK , ISMAÏL KHRISS , AND ABDESLAM JAKIMI
Abstract:	Software reverse engineering plays an important role when maintaining legacy systems, enabling understanding of a system by extracting high-level models from its source code. These models can represent the structural or behavioral aspects of the system. Several approaches have been proposed in the literature for recovering structural models, such as the Unified Modeling Language (UML) class diagram. Conversely, there is less work concerning extracting behavioral representations that capture different interactions within a given system. This paper investigates approaches to extracting behavioral UML diagrams, precisely sequence and use case diagrams. We have categorized these approaches into three groups, depending on the type of analysis employed: static, dynamic, or hybrid. Subsequently, we conducted a comparative analysis of these approaches, evaluating them based on various criteria to highlight their strengths and weaknesses. Based on this comparison, we propose an approach that combines static and dynamic analysis techniques to recover behavioral diagrams from source code. This proposed approach can potentially assist software developers in maintenance by providing a higher-level representation of a system that can even be employed in a modernization process to migrate it from a legacy environment to a modern one.
Keywords:	Reverse Engineering, Modernization Process, Behavioral model, UML Sequence Diagram, UML Use Case Diagram.
Source:	Journal of Theoretical and Applied Information Technology 30^th November 2023 -- Vol. 101. No. 22-- 2023
Full Text

Title:	UNLOCKING CHATGPT'S TITLE GENERATION POTENTIAL: AN INVESTIGATION OF SYNONYMS, READABILITY, AND INTRODUCTION-BASED TITLES
Author:	DUAA AKHTOM, OSAMAH MOHAMMED ALYASIRI, EMAN ALLOGMANI, AMER M. SALMAN, THAEER MUEEN SAHIB
Abstract:	The release of ChatGPT, an advanced language model by OpenAI, in November 2022 marked a significant milestone in Natural Language Processing (NLP). ChatGPT demonstrated exceptional capabilities in comprehending and generating human-like language, with applications spanning public health, climate research, education, academia, and more. One of its intriguing features is the generation of titles for various textual content, raising the question of whether ChatGPT can match or surpass human writers in title creation. This study investigates the efficacy of ChatGPT in creating titles for papers across the fields of Computer Science, Chemistry, Physics, and Medicine. The methodology employed examines title generation in scenarios both including and excluding synonyms. This approach aims to evaluate the effectiveness of the generated titles in terms of grammatical correctness, informativeness, conciseness, and engagement. Results reveal that considering synonyms leads to more diverse and contextually rich titles, whereas ignoring them yields titles that closely replicate the original terminology. The choice between these approaches depends on specific use cases. Additionally, variations in title lengths impact readability and SEO optimization, with some titles being shorter, others maintaining similar lengths, and some becoming more detailed. ChatGPT's potential to generate titles based on research paper introductions is also assessed, uncovering the role of synonyms in title variation. This research offers valuable insights into the capabilities and considerations when using ChatGPT for title generation.
Keywords:	ChatGPT, Natural Language Processing, Automated Title Generation, Readability, Engagement
Source:	Journal of Theoretical and Applied Information Technology 30^th November 2023 -- Vol. 101. No. 22-- 2023
Full Text

Title:	A DOCTOR RECOMMENDATION AND BREAST CANCER PREDICTION USING MODIFIED K-MEANS AND SVM
Author:	RETHINAKUMAR, GOPINATH GANAPATHY, JEONG-JIN KANG
Abstract:	Today, people are exposed to various diseases due to their living habits and environmental conditions. Early diagnosis is very important, but it can be hard to predict accurately due to the symptoms. The correct prediction is therefore the most challenging part of the job. Data mining is a process that can help predict a disease. Through the use of medical data, it can analyze the patterns of the disease. In this research the K-means algorithm is used in this study along with a support vector machine. In previous work, we utilized the CNN algorithm to identify breast cancer, but this was unsuitable for large-scale datasets. We have now modified this algorithm to handle training breast cancer information.The modified K-means algorithm can reduce the training number and produce a new dataset that is completely original. It also provides the necessary information about the predicted disease and its recommended doctors. The Proposed algorithm takes into account various factors such as the distance from the predicted location, the experience of the doctors, and the feedback of the users to provide a personalized recommendation. Through the proposed algorithm, the user can get the most effective and personalized treatment possible. It also reviews the recommended doctors and provides its own recommendations to improve the accuracy of the diagnosis.
Keywords:	K-means, Support Vector Machine, Data mining and Machine Learning
Source:	Journal of Theoretical and Applied Information Technology 30^th November 2023 -- Vol. 101. No. 22-- 2023
Full Text

Title:	FEATURES OF THE IMPORT SUBSTITUTION PROCEDURE IN THE CREATION OF UNMANNED AIRCRAFT VEHICLES TO INCREASE FLIGHT SAFETY
Author:	NATALYA V. PROSVIRINA , ALEXEY I. TIKHONOV
Abstract:	The topic of the study concerns the features of the import substitution procedure in the context of the creation of unmanned aerial vehicles (UAV) in order to improve flight safety. Import substitution is a strategically important mechanism for ensuring the country's independence in the development and production of technologically complex systems, such as UAV. The work examines the key stages of the import substitution procedure in the context of creating UAV and also identifies factors that contribute to increasing the level of flight safety. The incidents, which are currently the most important for the creation of unmanned aircraft systems (UAS) and their integration into a unified airspace, are analyzed: improving of flight safety and efficiency of target control tasks (piloting). The task of integrating unmanned aviation into a unified non-segregated airspace is largely due to its rapid development and widespread use. Integration is understood as the performance of flights by unmanned aviation on a regular basis together with manned aircraft in a single airspace with an acceptable level of flight safety. Considering that such regular flights of unmanned aircraft may threaten the safety of existing civil aviation, International Civil Aviation Organization (ICAO) has assumed the role of coordinator in creating all necessary conditions for the step-by-step solution of the integration problem. The article discusses the strategy for implementing the Global Plan for Flight Safety, as well as the phased plan for integrating drones into Russia's common airspace. The study also focuses on analyzing the technological and production aspects necessary for the successful implementation of import substitution in the field of UAV. Particular attention is paid to the selection of materials, technical characteristics and safety standards, as well as interaction with Russian suppliers of components and technologies. The results of the study can be used as recommendations for developers and manufacturers of UAV, as well as for the formation of strategies for government support for import substitution programs in the field of UAV, taking into account the priority of ensuring maximum flight safety.
Keywords:	Import Substitution, Import Independence, Flight Safety; Unmanned Aircraft; Remotely Piloted Aircraft (RPA); Safety Management System (SMS); Manual Control; Manned Aircraft; Aircraft Automation; Human Factor; Unmanned Aircraft System (UAS); Unified Airspace; Artificial Intelligence (AI); Aeromobility
Source:	Journal of Theoretical and Applied Information Technology 30^th November 2023 -- Vol. 101. No. 22-- 2023
Full Text

Title:	HOW CAN LIVE SHOPPING IMPROVE ACTUAL PURCHASE ON TIKTOK USERS?
Author:	MAULIDYA YUNIARTI ANWAR , VIANY TJHIN UTAMI
Abstract:	The purpose of this research is to determine the factors that affect TikTOk users in actual purchases through the use of live shopping or live commerce on the TikTok Shop. The structural model of this research includes various variables such as Trust, Perceived Enjoyment, Discount Framing, and E-Payment System that will be analyzed to have an effect on actual purchase actions by customers. The data will analyze using SEM-PLS with WarpPLS 7.0. From the result found that the Trust and Discount Framing had a significant effect on Actual Purchase direct or indirect through Purchase Intention. Meanwhile E-payment system only has significant effect on actual purchase. From thirteen hyphotesis tested, ten of them was accepted. The result of this study may assist TikTok Shop seller that used TikTok Live in evaluating and implementing features to increase actual purchase from their online store.
Keywords:	E-commerce, Live Shopping, TikTok, Actual Purchase, Purchase Intention.
Source:	Journal of Theoretical and Applied Information Technology 30^th November 2023 -- Vol. 101. No. 22-- 2023
Full Text

Title:	EFFICIENCY OF FAKE NEWS DETECTION WITH TEXT CLASSIFICATION USING NATURAL LANGUAGE PROCESSING
Author:	SALMAN AL FARISY AZHAR, FELIK HIDAYAT, MUHAMMAD HANIF AZFAREZAT, GHINAA ZAIN NABIILAH, ROJALI
Abstract:	Fake news has become a significant concern in today's information landscape, posing serious threats to society, democracy, and individual well-being. To combat the spread of fake news, effective detection mechanisms are essential. This paper investigates the efficiency of fake news detection through text classification using natural language processing (NLP) techniques. The study explores the application of various NLP algorithms, including feature extraction methods, sentiment analysis, and machine learning classifiers, to identify and classify news articles as either real or fake. The performance of different NLP approaches is evaluated using a comprehensive dataset comprising diverse news sources. In this paper we have used several methods of Algorithm such as Naïve Bayes, Random Forest, Logistic Regression, Decision Tree and, XGBoost. The results showed that certain algorithms like recurrent XGboost and Decision Tree machines performed well in detecting fake news with an accuracy score of 99.76% and 99.53%.
Keywords:	Fake news detection, text classification, NLP, machine learning classifiers, dataset. Random Forest, Logistic Regression, XGBoost, Decision Tree, Naive Bayes, Preprocessing
Source:	Journal of Theoretical and Applied Information Technology 30^th November 2023 -- Vol. 101. No. 22-- 2023
Full Text

Title:	AUTHENTICATED SECURITY SYSTEM BASED ON FREEHAND SKETCH USING FUZZY-WUZZY PARTIAL RATIO
Author:	N. KESAVA RAO, G. SRINIVAS, P.V.G.D. PRASAD REDDY, S. AMARNADH
Abstract:	This document presents the FreeHand Sketch-based Authentication Security System, a novel methodology for validation purposes. In the current digital world, security plays a significant role. The administrator takes many security measures, but still, it is getting hacked in different ways. To take digital data security to the next level, a new methodology is proposed with Fuzzy-Wuzzy with better accuracy. The approach allows users to register by sketching five similar images on their choice. These sketches undergo pre-processing using threshold with Gaussian mixture and Fuzzy-Wuzzy algorithm to assess their similarity. If all five pictures are deemed alike, they are stored in the database. For login, user can utilize their authorized information along with an image-based sketch password, which is also processed with threshold and Gaussian mixture and compared to the registered image passwords in the database for authentication using the Fuzzy-Wuzzy method. The proposed methodology's performance is evaluated using input sample image passwords and metrics like precision and recall. The proposed work demonstrates that user security can be ensured with an accuracy level of around 90% through authentication measures.
Keywords:	Biometric Systems, Authentication, Security, Fuzzy-Wuzzy, Authorization, Free Hand Sketch-Based Authentication, Security and Safety Patterns.
Source:	Journal of Theoretical and Applied Information Technology 30^th November 2023 -- Vol. 101. No. 22-- 2023
Full Text