		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

December 2023 | Vol. 101 No.23

Title:	DESIGN OF CONTEMPORARY MULTIVARIATE DATASET TO ASSESS THE QUALITY OF OBJECT, FACE AND PROXIMITY DETECTION IN ASSISTING THE VISUALLY IMPAIRED PEOPLE
Author:	S. SAJINI , Dr. B. PUSHPA
Abstract:	In recent years, advancements in Computer Vision have significantly impacted the development of assistive technologies for visually impaired individuals. The Major Objective of this research builds on the creation of a contemporary multivariate dataset designed to evaluate the quality of object, face, and proximity detection systems tailored to assist visually impaired individuals. The dataset incorporates diverse real-world scenarios, encompassing various environmental conditions and complexities commonly encountered by the visually impaired. It includes annotated images and accompanying ground truth data to facilitate the training and assessment of machine learning models for accurate object and face detection, as well as proximity estimation. The research work was based on the model designed using IoT enabled device and tested with 100 samples of visually impaired people. By leveraging this dataset, researchers and developers can enhance the performance of assistive technologies, ultimately improving the lives and independence of visually impaired individuals. The proposed dataset serves as a valuable resource for advancing the field of Computer Vision in the domain of accessibility and inclusive technology.
Keywords:	Object Detection; Face Detection; Proximity Detection; Multivariate Dataset; Visually Impaired; Computer Vision
Source:	Journal of Theoretical and Applied Information Technology 15^th Decmeber 2023 -- Vol. 101. No. 23-- 2023
Full Text

Title:	ACCELERATE DECISION-MAKING PROCESS THROUGH THE IMPLEMENTATION OF FINA: FINANCIAL FEASIBILITY STUDY IN HAND
Author:	WAHYU SARDJONO , ARVA BHAGAS
Abstract:	A financial feasibility study is sometimes a time-consuming process that makes management often misses opportunities or even make mistakes in making a decision. This paper describes the urgency of digitalization using an application or an automated tool that makes the investment decision process easier for everyone through dynamic financial analysis features, which is then called FiNA (financial feasibility study in hand). It aligns with the objective of Industry 4.0 and Indonesia Making 4.0 Agenda which requires digitization through continuous innovation. This paper explains the FiNA development process, how the application works in performing investment feasibility sensitivity calculations, and how to interpret the resulting information. However, FiNA as a tool that can accommodate many assumptions and considerations is still a long way to go and needs further development for improvement.
Keywords:	Feasibility Study, Financial Analysis, Decision Making, Fina, Industry 4.0
Source:	Journal of Theoretical and Applied Information Technology 15^th Decmeber 2023 -- Vol. 101. No. 23-- 2023
Full Text

Title:	AN IMPROVED MALWARE VARIANT DETECTION MODEL BASED ON HOMOGENEOUS STATIC HYBRID FEATURES AND A DATA AUGMENTATION TECHNIQUE
Author:	AZAABI CLETUS, ALEX AKWASI OPOKU, BENJAMIN ASUBAM WEYORI
Abstract:	The use of Machine learning has become the de-facto standard for malware defense due to the limitations of signature-based, heuristic-based and other cloud-based techniques. However, poor malware features, class imbalance problems and malware obfuscation remain challenges facing malware researchers. To ensure efficient and resilient detection in the face of these challenges requires novel models that adopt innovative techniques to improve malware detection. The paper proposed an improved novel malware variant detection model based on Homogeneous Multi-Static Hybrid features (HMSHF), obfuscated malware dataset and Synthetic Minority Oversampling Technique (SMOTE). A malware dataset comprising 11678 malware files from virusTotal.com and 3963 benign files obtained from windows environment was used for the study. We extracted ‘fine-grained’ strings, APIs, and opcode features from static disassembly of the malware dataset. We trained and tested a Random Forest (RF), Support Vector Machine (SVM), GradientBoost (GB), and eXtremGradientBoost (XGB) ensemble algorithms before and after obfuscating the malware dataset. We hybridized the features into HMSHF for training and testing the ensembles before and after the malware was obfuscated. We evaluated the performance of the models using individual features and the hybrid features before and after obfuscations. To overcome the class imbalance problem, we applied the SMOTE technique on the training set with the HMSHF. The proposed hybrid features showed effectiveness and efficiency in classifying malware with 99.87% accuracy without data augmentation and 98.8% accuracy with SMOTE data augmentation. Consequently, the paper concluded that, the proposed technique improved malware detection and demonstrated resilience against obfuscation compared with the state of the art. Thus, the approach can be adopted for the detection of known, unknown and zero-day malware. Notwithstanding the improved performance, this work is not without limitations; the use of feature selection instead of feature extraction, and use of ensembles instead of other Deep learning techniques and SMOTE instead of other data augmentation methods. Thus, future works will adopt the approach and use Principal Component Analysis (PCA) dimensionality reduction techniques; employ deep learning techniques and apply other data augmentation techniques to observe the performance.
Keywords:	SMOTE, Malware, Ensemble Learning, Ransomware, Malware Features, Signature-Based Detection
Source:	Journal of Theoretical and Applied Information Technology 15^th Decmeber 2023 -- Vol. 101. No. 23-- 2023
Full Text

Title:	DEMENTIA RISK ASSESSMENT USING MACHINE LEARNING AND PART-OF-SPEECH TAGS
Author:	AKSHAY ZADGAONKAR, RAVINDRA KESKAR, OMPRAKASH KAKDE
Abstract:	Dementia, a set of cognitive decline syndromes distinct from typical age-related degeneration, poses a significant public health challenge. The key to dementia detection lies in analyzing sentence structure and conversational style, particularly in speech. This study focuses on creating and evaluating a machine learning model for non-invasive early dementia detection through speech parameter analysis in everyday conversation. Leveraging the DementiaBank dataset, comprising over 500 voice transcripts from individuals aged 60 and older, the study employs 63 tagged Part-of-Speech (PoS) parameters extracted from chat transcripts. Data from 244 control subjects and 306 dementia patients are used. Machine learning methods, including Random Forest, Deep Neural Network, and Support Vector Machine, achieve respective accuracy rates of 83%, 92%, and 84%. These results underscore the effectiveness of informatics-based machine learning in non-invasive dementia detection using PoS tags. Additionally, the study provides insights into the relative importance of each PoS tag in dementia detection. This research contributes to the growing informatics field of dementia detection and supports the development of less intrusive diagnostic tools.
Keywords:	Machine Learning, Dementia, Speech, Linguistics, Part-of-speech
Source:	Journal of Theoretical and Applied Information Technology 15^th Decmeber 2023 -- Vol. 101. No. 23-- 2023
Full Text

Title:	REVOLUTIONIZING COMPUTER VISION: ENHANCED FOOD IMAGE CLASSIFICATION WITH SWIN TRANSFORMER AND SVM CLASSIFIER
Author:	ELPINA, GEDE PUTRA KUSUMA
Abstract:	This paper presents a novel approach for food image classification using a combination of the Swin Transformer model and a support vector machine (SVM) classifier. The proposed method surpasses the performance of the original Swin Transformer model trained on ImageNet, achieving an impressive accuracy of 91.05% on the testing dataset. Comparative evaluation shows that the SVM classifier enhances the classification capabilities of the Swin Transformer, outperforming the baseline approach. The results highlight the efficacy of the Swin Transformer as a feature extraction model for food image classification tasks. The integration of deep learning with traditional machine learning techniques, as demonstrated by the SVM classifier, shows promise for improving classification accuracy in various applications such as food recognition systems and dietary analysis tools. Future work includes further optimization of the proposed method, exploring domain adaptation and transfer learning techniques, and investigating advanced fusion methods to achieve even higher classification accuracy and improved generalization across diverse food domains.
Keywords:	Food Image Classification, Deep Learning, Vision Transformer, Swin Transformer, Feature Extraction, Support Vector Machine
Source:	Journal of Theoretical and Applied Information Technology 15^th Decmeber 2023 -- Vol. 101. No. 23-- 2023
Full Text

Title:	NOVEL SIGNATURE SCHEMES FOR MULTI-MESSAGE SIGNING WITH A SINGLE PUBLIC KEY USING POST-QUANTUM DIGITAL SIGNATURE ALGORITHMS IN MANET
Author:	R.PRIYAVANI, DR.N.KOWSALYA
Abstract:	Mobile Ad hoc Networks (MANETs) are decentralized networks that organize themselves, forming multi-hop connections through an ever-changing and unpredictable topology. In this context, any node can function as a sender, receiver, or router, facilitating peer-to-peer communication without relying on centralized infrastructure. Given the reliance on battery power for mobile nodes, the instantaneous connectivity of diverse devices within this network can sometimes result in instances of non-acknowledgement behavior, potentially causing network performance deterioration. To mitigate this performance degradation, we propose an innovative approach using Post-Quantum Cryptography (PQC) tailored to handle non-acknowledgement data in MANETs. Specifically, our method incorporates a customizable hash function called "Everything tweak able hash function" to establish a reliable end-to-end solution. This solution introduces location-aware post-quantum encryption, effectively countering non-acknowledgement data behavior within a bi-directional multi-hop relay setup. Our novel Post-Quantum Cryptography (NPQC) algorithm not only focuses on addressing non-acknowledgement data concerns but also seeks to comprehensively assess the implications of this issue. By evaluating key metrics such as key generation time, encryption/decryption time, security level, execution time, and memory consumption, our aim is to achieve notable enhancements in execution time and overall security within the dynamic environment of an MANET.
Keywords:	Post Quantum Cryptography, Multivariate Cryptography, Hash Based Signature, Tweakable Hash Function, MANET
Source:	Journal of Theoretical and Applied Information Technology 15^th Decmeber 2023 -- Vol. 101. No. 23-- 2023
Full Text

Title:	COMPARATIVE ANALYSIS OF PREDICTIVE MODELS FOR WORKLOAD SCALING IN IAAS CLOUDS: A STUDY ON MODEL EFFECTIVENESS AND ADAPTABILITY
Author:	SATYA NAGAMANI POTHU,DR. SWATHI KAILASAM
Abstract:	The demand for dependable workload prediction models has surged in the ever-evolving domain of cloud computing, especially across renowned platforms such as AWS, Google Cloud, and Azure. These models are instrumental in enabling efficient resource allocation and enhancing overall performance. This comparative research focuses on various predictive models pivotal for reactive and proactive scaling in Infrastructure as a Service (IaaS) clouds. Initially, the study evaluates time series and machine learning models. These models have shown prowess in accurately forecasting workloads on real-time cloud datasets, leading to notable savings in resource allocation. However, their effectiveness can be challenged during abrupt changes in workload, underscoring the need for more dynamic modeling approaches. The research then delves deeper into Markov models and their simulations on real-time cloud datasets. These models, rooted in state transitions and probabilistic events, have been a cornerstone in predicting resource demands and optimizing workload distribution in cloud environments. Simulations based on Markov models provide valuable insights into potential future states, making them an invaluable tool for proactive resource management. Nevertheless, the intricacies involved in these simulations, especially when handling large-scale real-time datasets, can sometimes act as a double-edged sword, leading to computational challenges and necessitating further optimization. The study also touches upon reinforcement learning models, which have been significant in resource management and performance enhancement. However, these models come with their challenges, where the complexity of their learning algorithms might sometimes hinder optimal performance. This observation paves the way for a recommendation to refine and streamline the learning processes to bolster their efficiency. The research concludes with an examination of evidence-based design and simulation models. While adept at assessing specialized aspects, such as visual comfort in modern office designs, their performance can be compromised by the complexities associated with their simulation methods. The specific use case and inherent requirements influence the ideal predictive model. While particular models excel in more stable settings, others are tailored for unpredictable environments. The future beckons a focus on refining these models, ensuring they are well-equipped to handle abrupt changes and the multifaceted nature of cloud settings, thereby maximizing the potential of cloud computing services.
Keywords:	Predictive Models, Workload Prediction, Reactive Scaling, Proactive Scaling, IaaS Clouds, Machine Learning Models.
Source:	Journal of Theoretical and Applied Information Technology 15^th Decmeber 2023 -- Vol. 101. No. 23-- 2023
Full Text

Title:	DETECTION OF FACIAL MICRO-EXPRESSIONS USING CNN
Author:	F. TADEO SÁNCHEZ GARCÍA, B. LÓPEZ LIN, RODOLFO ROMERO-HERRERA
Abstract:	Brief, involuntary micro-facial expressions represent a window into a persons hidden or repressed emotions. The ability to analyze and detect them can have a significant impact in various fields that require an understanding of human behavior. However, the process of detecting micro-expressions poses significant challenges, such as the implementation of the detection method or the generation of extensive and quality data. This article develops a Machine Learning model with a convolutional neural network; It is compared with other existing models of micro expressions in the prediction of one of 7 human emotions, to recognize micro expressions and predict one of 7 emotions.
Keywords:	Deep Learning, Human Emotions, Micro-Expressions, Convolutional Neural Networks.
Source:	Journal of Theoretical and Applied Information Technology 15^th Decmeber 2023 -- Vol. 101. No. 23-- 2023
Full Text

Title:	AN EXTENSION TO THE TRIANGULATION TECHNIQUE FOR EFFORT ESTIMATION IN SOFTWARE DEVELOPMENT PROCESS
Author:	RAID ZAGHAL‎‎, ODAY AL-WAHSH‎, SAEED SALAH‎, HANA SHABANEH
Abstract:	Effort Estimation (EE) is one way to evaluate a software project and understand its schedule and budget. It is one of the most important aspects of the software development life cycle. The failure of not recognizing the accurate effort estimation may lead to increase the financial costs of the companies and their clients which will cause negative impact on their job duties and their future marketing plans besides the client’s disappointment and dissatisfaction. The Palestinian IT sector is one of the most developing and promising sectors. However, the studies that investigate the methods and techniques of effort estimation are most likely missing. For that reason, we were motivated to study the status of the software development companies in Palestine to better understand how the technical teams estimate the needed effort of their software projects. The purpose of this study is (1) to survey the existing effort estimation techniques used by prominent Palestinian software development companies and analyze their practices, and (2) to suggest an appropriate effort estimation technique that can suite the nature and needs of these companies and to validate this technique via real application within actual software projects in a selected subset of these companies. Based on our survey and analysis, we have selected an existing effort estimation technique called (Triangulation) as the most appropriate EE method for small companies and for the Palestinian software development industry. After that, we have designed an extension of this technique to (fine-tune) the company’s exact needs and provide it with some flexibility to make some changes and adjustments (e.g., decrease project delivery time by increasing resources). Moreover, this technique was applied on the Palestinian software projects to validate its results. We believe this study can be a valuable resource for Palestinian software development companies; as they can use it as a guideline to help them get better and more accurate effort estimates, which in return can reduce costs and provide better and more accurate scheduling and staffing needs.
Keywords:	Software Development, Effort Estimation Technique, Ensemble Effort Estimation, Software Engineering, Agile
Source:	Journal of Theoretical and Applied Information Technology 15^th Decmeber 2023 -- Vol. 101. No. 23-- 2023
Full Text

Title:	LEARNING FROM DISASTER: DISASTER RELIEF MANAGEMENT USING DEEP LEARNING
Author:	NEKURI NAVEEN, DELTON M ANTONY1, RAGHAVA M
Abstract:	Our primary contributions in this article are the development of deep learning models for disaster management. The research work proposes three architectures with two variants each: A Long short-term Memory (LSTM) based model, a Convolutional Neural Network (CNN) based model, and a CNN-LSTM based model. Each of these architectures will have two variants built. The first variant is the one where the word embeddings are learned from the supervised data itself. The second one is where the word embeddings are pretrained. The results are presented based on empirical evidence and conclusions are highlighted.
Keywords:	Sentiment Analysis, CNN, LSTM, Multi-Model, Word Embeddings, Classification
Source:	Journal of Theoretical and Applied Information Technology 15^th Decmeber 2023 -- Vol. 101. No. 23-- 2023
Full Text

Title:	THE INTERNET OF VEHICLES (IOV) TECHNOLOGY: CHALLENGES AND SOLUTIONS
Author:	ANAS S. ALKARIM, ABDULLAH S. AL-MALAISE AL-GHAMDI, MAHMOUD RAGAB
Abstract:	The Internet of Things (IoT) revolution has paved the way for the emergence of Internet of Vehicles (IoV) technology, enabling seamless communication and data exchange among vehicles, infrastructure, and pedestrians. This paper delves into the IoV landscape, examining its challenges, solutions, and the role of artificial intelligence (AI) methods in addressing critical issues. The paper begins by elucidating the foundational concepts of IoV, emphasizing its potential to revolutionize transportation systems through communication protocols, Vehicle-to-Everything (V2X) technology, cybersecurity, data management, edge computing, and artificial intelligence (AI). However, realizing these benefits involves numerous challenges, including managing massive amounts of data, addressing data privacy and security concerns, mitigating network congestion, ensuring reliability, and achieving scalability. This paper comprehensively analyses IoV technology, explores the associated challenges, and presents innovative solutions enabled by artificial intelligence. By harnessing the potential of AI methods, the IoV ecosystem can evolve into a safer, more efficient, and sustainable transportation paradigm, revolutionizing how we navigate and interact with urban environments.
Keywords:	The Internet of Vehicles, Intelligent Transportation Systems, Artificial Intelligence, Machine Learning, Smart Cities.
Source:	Journal of Theoretical and Applied Information Technology 15^th Decmeber 2023 -- Vol. 101. No. 23-- 2023
Full Text

Title:	FROM CONVENTIONAL TO DIGITAL MEDIA: DIGITAL TRANSFORMATION STRATEGIES ON METRO TV IN INDONESIA
Author:	AFDAL MAKKURAGA PUTRA, ANDI SETIA GUNAWAN, NOVI ERLITA
Abstract:	The development of technology and the internet has changed how society consumes media. Technological developments disrupt conventional media and carry out digital transformation in response to changing conditions. Metro TV only responded to this change in March 2022 by establishing Digital Hub as the central kitchen for the digitization process. This study aims to find out the background, process, and form of changes produced by Digital Hub Metro TV. The method used is a case study with a qualitative descriptive approach—primary data collection through in-depth observation and interviews of 4 informants and key informants. Literature reviews, corporate records, and other secondary sources were used to gather secondary data. Discussion of research results using The Long Tail theory as a basis for researchers in analysing data. The results showed: 1) Metro TV was late in digital transformation because it initially saw social media as a "threat" to mainstream media. 2) Metro TV makes Digital Hub the central kitchen of the transformation process to improve quality, extend versions, and diversify content to social media. 3) Metro TV applies the principle of The Long Tails three forces: production democratization, distribution democratization, and supply and demand connection.
Keywords:	Digital Transformation, Long Tail Economics, Television, Digitalization.
Source:	Journal of Theoretical and Applied Information Technology 15^th Decmeber 2023 -- Vol. 101. No. 23-- 2023
Full Text

Title:	SELF-ADAPTIVE LION PRIDE OPTIMIZATION-BASED ENHANCED RANDOM FOREST (SALPO-ERF) ALGORITHM FOR ENHANCED TRAFFIC SURVEILLANCE OBJECT DETECTION
Author:	VALARMATHI V, Dr.S.DHANALAKSHMI
Abstract:	Traffic surveillance is crucial for modern urban infrastructure, providing real-time data on traffic conditions, vehicle movements, and road safety. It aids traffic management, accident prevention, and law enforcement. Ensuring road safety is a top priority, and traffic surveillance significantly reduces accidents. An enduring challenge in traffic surveillance is accurate object detection under diverse conditions, including adverse weather, low lighting, and occlusions. Traditional algorithms often struggle in these scenarios, potentially jeopardizing safety and traffic management. The Self-Adaptive Lion Pride Optimization-based Enhanced Random Forest (SALPO-ERF) algorithm is developed to address the challenges in object detection in traffic surveillance. SALPO-ERF combines SALPO’s adaptability with ERF’s enhanced object detection capabilities. SALPO adjusts feature importance dynamically, even in complex and noisy traffic situations, while ERF provides a strong foundation for robust object detection. SALPO-ERF’s evaluation on the AAU RainSnow Traffic Surveillance Dataset, featuring 22 five-minute videos and 13,297 objects, demonstrated superior object detection accuracy and robustness compared to state-of-the-art algorithms. This underscores SALPO-ERF’s potential to significantly enhance traffic surveillance accuracy and contribute to safer, more efficient roadways.
Keywords:	Optimization, Object Detection, Traffic Surveillance, Random Forest, Lion Pride, Fitness
Source:	Journal of Theoretical and Applied Information Technology 15^th Decmeber 2023 -- Vol. 101. No. 23-- 2023
Full Text

Title:	AN IMPROVED REAL-TIME HANDGUN DETECTION SYSTEM USING YOLO V5 ON A NOVEL DATASET
Author:	IMANE RAHIL, WALID BOUARIFI, RAHIL GHIZLANE, OUJAOURA MUSTAPHA
Abstract:	In the face of widespread gun violence, it has become imperative to enhance the capabilities of public surveillance cameras by integrating intelligent automatic handgun detection systems. This study presents a comprehensive approach to automate the real-time identification of pistols in video security footage using the advanced YOLO-V5 algorithm. A carefully curated dataset of varied pistol images was employed to optimize the models performance across diverse scenarios and minimize dependence on human security personnel. Recognizing the crucial implications of firearm detection in images for public safety and law enforcement, this study employed advanced techniques such as data augmentation, transfer learning, and test time augmentation to enhance the models performance. Iterative fine-tuning of hyperparameters was conducted to attain the desired level of accuracy. The results demonstrate that the YOLO-V5 model exhibits high precision and recall in detecting handguns, even in complex and challenging environments. This study represents a significant advancement in the development of effective gun detection systems, serving as a catalyst for further research in this exciting field. By automating the identification of firearms in real-time video surveillance, this approach addresses a critical need for enhanced public safety measures and offers valuable insights into the potential of intelligent surveillance technologies
Keywords:	Yolo v5, Handgun Detection, Machine Learning, False Positive, Guns, Computer vision, Transfer Learning, Deep Learning, Violence Detection, Faster R-CNN.
Source:	Journal of Theoretical and Applied Information Technology 15^th Decmeber 2023 -- Vol. 101. No. 23-- 2023
Full Text

Title:	DETECTION OF FACIAL MICRO-EXPRESSIONS USING CNN
Author:	F. TADEO SÁNCHEZ GARCÍA, B. LÓPEZ LIN, RODOLFO ROMERO-HERRERA
Abstract:	Brief, involuntary micro-facial expressions represent a window into a persons hidden or repressed emotions. The ability to analyze and detect them can have a significant impact in various fields that require an understanding of human behavior. However, the process of detecting micro-expressions poses significant challenges, such as the implementation of the detection method or the generation of extensive and quality data. This article develops a Machine Learning model with a convolutional neural network; It is compared with other existing models of micro expressions in the prediction of one of 7 human emotions, to recognize micro expressions and predict one of 7 emotions.
Keywords:	Deep Learning, Human Emotions, Micro-Expressions, Convolutional Neural Networks.
Source:	Journal of Theoretical and Applied Information Technology 15^th Decmeber 2023 -- Vol. 101. No. 23-- 2023
Full Text

Title:	HARMONY SEARCH OPTIMIZATION-BASED GENERATIVE ADVERSARIAL NETWORKS (HSO-LGAN) FOR ENHANCING SENTIMENT ANALYSIS IN AUGMENTED REALITY-ENABLED APPLICATIONS
Author:	PRAGTHI ARAVABOOMI, T.K.SARAVANAN
Abstract:	This paper explores the transformative dynamics of online shopping, emphasizing the pivotal role of augmented reality (AR) in reshaping the customer experience. In the digital marketplace, customer product reviews serve as a rich source of feedback, providing valuable insights for businesses. However, the nuanced sentiments within these reviews pose a significant challenge for effective analysis. Acknowledging this complexity, the proposed work introduces "Harmony Search Optimization-Based Generative Adversarial Networks (HSO-LGAN)" as an innovative solution. HSO-LGAN leverages harmony search optimization algorithms within generative adversarial networks to translate textual sentiments from product reviews into immersive AR experiences. The working mechanism involves a synergistic approach, enhancing the generative adversarial networks capabilities to create personalized and emotionally resonant AR content. Through comprehensive evaluation, the paper assesses the effectiveness of HSO-LGAN in improving customer engagement, satisfaction, and emotional attachment to brands. Results and discussions from experimental studies across diverse industries highlight the potential of HSO-LGAN to significantly impact customer loyalty and advocacy in the evolving landscape of online shopping. This research contributes to the broader understanding of leveraging AR and sentiment analysis to optimize the online shopping experience and bridge the gap between customer feedback and interactive AR elements.
Keywords:	Augmented Reality, Classification, Gan, Harmony Search, Sentiment Analysis, Optimization.
Source:	Journal of Theoretical and Applied Information Technology 15^th Decmeber 2023 -- Vol. 101. No. 23-- 2023
Full Text

Title:	RANDOM WALK-BASED APPROACH FOR ADDRESSING OF NAVIGATION ALGORITHMS IN SERVICE AND SURVEILLANCE ROBOTS
Author:	ALBERTO MALDONADO ROMO, JESUS YALJA MONTIEL PEREZ, CESAR CASTREJON PERALTA, LUIS ENRIQUE ANDRADE GORJOUX, JOSE ALBERTO TORRES LEON
Abstract:	In the present article, a simulation of a robotics application based on random walk is described. Typically, robots employ navigation algorithms based on planning and adjustments with respect to their working environment. As a specific case for mobile robots with two-dimensional navigation, the navigation algorithm is implemented with the characteristics of a random walk. The route planning module is a stochastic process in which it operates with random numbers following a uniform distribution as input data to the navigation module, and the outputs are movement instructions, both in direction and distance of advance. As a test, a closed scenario is used with 1, 2, and up to 5 robots operating simultaneously. The objective is to cover the area of the scenario. The results include the percentage of visited area, as well as the decision metrics of the stochastic processes involved in navigation. This navigation algorithm is oriented towards applications in service and surveillance robots.
Keywords:	Random Walk, Stochastic Processes, Simulation, Robotics, Random Numbers.
Source:	Journal of Theoretical and Applied Information Technology 15^th Decmeber 2023 -- Vol. 101. No. 23-- 2023
Full Text

Title:	UNDERSTANDING CUSTOMER SATISFACTION AND LOYALTY TOWARD FOOD DELIVERY APPLICATION THROUGH USES GRATIFICATION APPROACH: MEDIATED BY CUSTOMER TRUST
Author:	WANDA WANDOKO, IGNATIUS ENDA PANGGATI
Abstract:	The aim of this research is to test the customer trust variable as a mediation between customer satisfaction and customer loyalty at food delivery application or FDA. Another aim of this research is to examine the influence of customer experience, delivery experience, navigation design and information design on customer satisfaction using paradigm of uses gratification theory. Respondent data from the research consisted of 600 respondents who used FDA in Indonesia. This research uses SMART PLS to test the research model. The results of the research show that the customer trust plays a significant role in partially mediating the influence of customer satisfaction on customer loyalty. Then delivery experience, customer experience, navigation design and information design have a positive influence in forming customer satisfaction. This research conducts empirical tests on the role of customer trust as a mediator in customer satisfaction and customer loyalty relationships in the context of food delivery applications. This research also investigate the influence of customer experience and delivery experience on their satisfaction. This research has theoretical implications for filling the gap in the literature on uses gratification theory and customer loyalty at the FDA.This research have several implications for FDA companies.
Keywords:	Trust; Food Delivery Application; Loyalty; Navigation Design; Information Design
Source:	Journal of Theoretical and Applied Information Technology 15^th Decmeber 2023 -- Vol. 101. No. 23-- 2023
Full Text

Title:	UNVEILING THE EVOLUTION: HOW HISTORY, POLITICS, CULTURE, AND TECHNOLOGY SHAPE ACCOUNTING SYSTEMS FOR SMEs IN INDONESIA
Author:	HARIADI YUTANTO, ROMI ILHAM, CANDRANINGRAT, ROHMAD FUAD ARMANSYAH
Abstract:	The historical documentation in Indonesia highlights the significant role of Micro, Small, and Medium Enterprises (MSMEs) in fostering economic development. This project aims to develop an integrated accounting information system application, known as SMESH Platform - Small Medium Enterprise Headquarters, for the Medokan Ayu Village Community Empowerment Institute (LPMK) in Surabaya. LPMK currently accommodates 54 MSMEs. The proposed system incorporates digital marketing functionalities inside the accounting information system. MSMEs frequently encounter challenges related to administrative tasks, leading to a significant allocation of time towards bookkeeping activities. Consequently, this diversion of resources may result in decreased production levels. However, MSMEs can potentially mitigate this issue by prioritising digital marketing efforts, thereby optimising their overall operational efficiency. This study involves conducting practical research on the design of information systems for SMEs using the System Development Life Cycle (SDLC) approach, specifically employing the Waterfall technique. The research encompasses seven stages: planning, analysis, design, development, testing, deployment, and maintenance. The present study focuses on the development of an integrated accounting information system named SMESH (Small Medium Enterprise Headquarters) that incorporates digital marketing strategies.
Keywords:	SMEs; Accounting Information Systems; Digital Marketing; Website.
Source:	Journal of Theoretical and Applied Information Technology 15^th Decmeber 2023 -- Vol. 101. No. 23-- 2023
Full Text

Title:	MODIFIED RANDOM FOREST REGRESSION MODEL FOR PREDICTING WHOLESALE RICE PRICES
Author:	CHRISTINE DEWI, GRUDA SAKTI KRIDA PRASATYA, HENOCH JULI CHRISTANTO, SANDRA OCTAVIANI B WIDIARTO4, GUOWEI DAI
Abstract:	Both in terms of diet and economy, Indonesian people attach great importance to rice as a staple food. In addition, it is very important to monitor rice price fluctuations every month so that overall rice prices remain stable and do not burden the community. Tracking rice price fluctuations helps rice producers, traders, and businesses make informed decisions about when to buy, sell, or store rice. This can optimize their supply chain management and maximize profits. Researchers and analysts can use rice price data to study market trends, identify patterns, and develop predictive models for future price movements. This research purpose to determine the most optimal forecasting model by using the Average Rice Price dataset at the Indonesian Wholesale Trade Level from January 2010 to December 2022. The dataset is obtained from the Central Statistics Agency of Indonesia. Moreover, the best model proposed in this research uses the Random Forest method with hyperparameter tuning using the n estimator parameter of 500. Our proposed method can reduce the MAPE value from 0.0093573 to 0.0089389 and increase the R2 Score value from 0.9916805 to 0.9921578. Moreover, we analyze the performance of our proposed methodology with several other datasets sourced from UCI (University of California Irvine). The experimental outcomes indicate that the suggested model displays superior performance when compared to alternative methods, with a tendency of decreasing MAPE values and increasing R2 values in each experiment for all datasets.
Keywords:	Random Forest, Rice Price Prediction, Machine Learning.
Source:	Journal of Theoretical and Applied Information Technology 15^th Decmeber 2023 -- Vol. 101. No. 23-- 2023
Full Text

Title:	DEVELOPMENT OF AI PREDICTIVE MODEL FOR MATHEMATICS LEARNING ACHIEVEMENT USING DEEP LEARNING
Author:	YOUNGHO LEE
Abstract:	In this study, a model for predicting learners mathematics achievement was developed using deep learning technology. It was using a data set of mathematics characteristics and achievement data of first-year elementary school learners in Korea. Using this data, in this study, we developed an artificial intelligence model that predicts previous or subsequent performance based on the current mathematical performance. In this study, it is vital to consider the mathematical characteristics of learners to predict learning achievement. For this, cluster analysis was conducted on initial mathematical functions (number size, number order, number counting), computational fluency, and cognitive processing (work memory, processing speed). Next, based on the results of mathematics learning achievement, we developed an artificial intelligence model that can predict mathematics achievement before or after. The artificial intelligence model was developed using a sequence-to-sequence (seq2seq) model of a recursive neural network (RNN) method to use continuous data as input/output. The model predicting the achievement of Unit 3 <Addition and Subtraction> with the achievement of Unit 1 <Numbers up to 9> in the 1st grade showed more than 90% accuracy and more than 98% recall rate.
Keywords:	AI Model, Learning Achievement Prediction, Artificial Intelligence, RNN, Seq2Seq, Edutech
Source:	Journal of Theoretical and Applied Information Technology 15^th Decmeber 2023 -- Vol. 101. No. 23-- 2023
Full Text

Title:	SALINE FLUID FLOW SUPERVISION IN INTENSIVE CARE UNIT USING PRECISION ALGORITHM
Author:	K.VIDYA SAGAR, SURYA PRASADA RAO BORRA, A. GEETHA DEVI, M. RAJ KUMAR NAIK, LAKSHMI RAMANI BURRA, VEERA VASANTHA RAO BATTULA, TATA BALAJI
Abstract:	Saline fluid infusion in intra vein supervision is imperative to correct significant abnormalities. The fluid infusion rate influence on the physiological parameters like systolic blood pressure (SBP), diastolic blood pressure (DBP), heart rate (HR), pulse rate (PR) and respiration rate (RR). The temperature (T) will change with rapid infusion of saline fluid. Saline fluid level monitoring is also significant. Monitoring these parameters continuously using Atmega 328 controller and energizing the alarm system and sending alert signals to the registers persons to curtail further disorders. With preset volumetric flow rate of the fluid patient is observed for 30 minutes with dextrose solution. Heart rate and cardiac stroke volume (CSV) and cardiac stroke volume index (CSVI) values change with increasing fluid infusion rate. Ther saline fluid infusion rate is controlled based on the changes of HR level, CSV, CSVI, SBP, and DBP variations, body temperature and RR values. The experimental finding shows blood pressure variation (BPV) change of 11.56%. systolic blood pressure (SBP) is increased 8% and diastolic blood pressure is decreased 3.25%. The heart rate change is 1.25 BPM. After 30 minutes of observation the heart rate change is 12%. This is significant.
Keywords:	Saline Fluid Monitoring, Systolic Blood Pressure, Diastolic Blood Pressure, Heart Rate, Pulse Rate, Cardiac Output, Cardiac Stroke Volume.
Source:	Journal of Theoretical and Applied Information Technology 15^th Decmeber 2023 -- Vol. 101. No. 23-- 2023
Full Text

Title:	A COLLABORATIVE QUANTUM ASSISTED EXTENDED ELLIPTIC CURVE CRYPTOGRAPHY TECHNIQUE FOR SECURE DATA TRANSMISSION OVER NETWORK
Author:	GOVINDU SURLA, R. LAKSHMI
Abstract:	The notion of a quantum computer is no longer just theoretical. It is the most significant technology in the world, and nations are competing to become the leaders in quantum computing. The computing time will be cut down from years to hours or even minutes thanks to technology. The scientific community will greatly benefit from the capabilities of quantum computing. It does, however, represent severe risks to cyber security. All encryption algorithms are theoretically prone to damage. Compared to RSA-based cryptosystems, elliptic curve cryptography (ECC) is quicker, more effective, and more sensitive to quantum attacks. Standard ECC is still unworthy of establishing a secure network connection, nonetheless. The improved ECC method is used to extend the communication strategy, reconfiguring the message with the number of cipher-text from both sides. Therefore, we need to carefully evaluate the quantum security of EECC to prepare for the advent of quantum computers. In this, work a new strategy (CQAEECC) known as a collaborative quantum-assisted Extended Elliptic Curve Cryptography (EECC) to protect the transmission of information across networks. The mechanism of merging cryptographic methods and the private key is retrieved from the Quantum Cryptography used by Extended Elliptic Curve Cryptography to ensure greater security over networks. The novel cryptography is compared with standard algorithms and the results show that it is one of the most efficient public key cryptosystems (PKC) for desirable security. Thereupon, the proposed method has the ability to ensure confidentiality, integrity, and availability over the network.
Keywords:	Quantum Computing, Public Key Cryptography, Elliptic Curve Cryptography
Source:	Journal of Theoretical and Applied Information Technology 15^th Decmeber 2023 -- Vol. 101. No. 23-- 2023
Full Text

Title:	TOWARDS A NEW STUDENTS SATISFACTION MODEL IN ONLINE ENVIRONEMENTS FOR ACADEMIC PERFORMANCE
Author:	MERYEM ZIZ , SOPHIA FARIS , KHALIFA MANSOURI
Abstract:	Online learning has established itself as a significant component of education, and models of user satisfaction and continuity of use are critical for assessing the efficiency of online learning platforms. This paper reviews prominent models such as the Technology acceptance model (TAM), The expectation confirmation model (ECM), Delone and Mclean information system success model (D&M ISS), Self-regulated learning (SRL), Task-technology fit (TTF), Self-Determination Theory (SDT), The innovation diffusion theory (IDT), Social Cognitive Theory (SCT), and Community of Inquiry (CoI), these theories have been developed to measure factors that leads to success in the online learning environment. Our study not only criticizes these models but aims to make a significant contribution to the field. The research contribution of this study is to propose a new model of learner satisfaction that leads to improve academic performance, this model is created based on the combination of several models such as CoI, SDT, TAM, ECM, IDT, SRL, TTF, D&M ISS , our model offers a comprehensive understanding of student behavior that allow online learning providers to better understand students’ behavior and make defensible judgments to enhance their platforms and boost user engagement. Our research stands as a significant contribution, advancing a more holistic understanding of student satisfaction and contributing to the ongoing evolution of online education platforms.
Keywords:	E-Learning; Models Of User Satisfaction; Continuity Of Use; Performance Impact
Source:	Journal of Theoretical and Applied Information Technology 15^th Decmeber 2023 -- Vol. 101. No. 23-- 2023
Full Text

Title:	STOCK MARKET PREDICITION USING STATISTICAL & DEEP LEARNING TECHNIQUES
Author:	MARAM GHIDAN ALQAHTANI, HODA AHMED ABDELHAFEZ
Abstract:	Predicting stock marketing prices has persistently a challenge due to the complexity of the stock data. Accurately predicting a stock s short-term price can increase the rate of investment and business opportunities in the stock market. This study aims to predict the closing prices of six major sectors in the Saudi stock market: Banking, Basic Materials, Real Estate Management and Development, Insurance, Energy, and Telecommunication. The dataset was historical records of the six sectors for seven years, along with two economic indicators: oil prices and inflation rates. Six models were employed for prediction: Auto-Regressive Integrated Moving Average (ARIMA), Support Vector Regression (SVR), Random Forests, Long Short-Term Memory (LSTM), Bidirectional Long Short-Term Memory (Bi-LSTM), and gated recurrent units (GRU). The models were evaluated using four regression metrics: mean squared error (MSE), mean absolute error (MAE), the mean absolute percentage error (MAPE), and the root mean squared error (RMSE). The findings revealed that GRU and Random Forests exhibit superior performance across multiple sectors, while SVR and Bi-LSTM demonstrated promising results. However, ARIMA consistently performed poorly across all sectors. The study provided valuable insights into the effectiveness of different models in predicting stock prices in the Saudi stock market. These findings could aid investors, analysts, and decision-makers in making informed investment decisions.
Keywords:	Stock Market Prediction; Machine Learning; AREMA; Deep Learning Techniques.
Source:	Journal of Theoretical and Applied Information Technology 15^th Decmeber 2023 -- Vol. 101. No. 23-- 2023
Full Text

Title:	PERFORMANCE AND SCALABILITY OF IPV4/IPV6 TRANSITION MECHANISMS FOR REAL-TIME APPLICATIONS
Author:	KHALID EL KHADIRI, NAJIB EL KAMOUN, SAMIR EL OUAHAM, OUIDAD LABOUIDYA, KAWTAR SMAHI, RACHID HILAL
Abstract:	To access the Internet, every device requires an IP address. However, the number of available IPv4 addresses is limited and insufficient to meet the growing demand for new addresses for a multitude of connected devices, including IoT devices and smartphones. In February 2011, the Internet Assigned Numbers Authority (IANA) announced the exhaustion of the /8 blocks of IPv4 addresses allocated to Regional Internet Registries (RIRs). Subsequently, the RIRs themselves exhausted their address reserves. Therefore, it is imperative to deploy the new version of the Internet Protocol, namely IPv6, which offers a significant expansion of the available address space. However, due to the incompatibility between IPv4 and IPv6, given their different headers, the transition from the old version (IPv4) to the new version (IPv6) cannot be achieved in a short period of time, requiring a gradual deployment. To address this challenge, three solutions are possible: a) Equip each device with a dual stack of IP addresses, b) Use tunneling to route IPv6 packets through the existing IPv4 network and c) Implement IP address translation, Among these options, tunneling is generally considered the most viable solution. However, it is worth noting that, like any technology, tunneling is influenced by potential scalability issues that need to be considered and managed to ensure a successful large-scale transition from IPv4 to IPv6. This article presents a comprehensive experimental study of the performance and scalability of IPv4 to IPv6 transition mechanisms. Our research is based on practical implementation in the GNS3 environment, where we increased the number of clients and explored various transition technologies to determine the most scalable solution. To evaluate these mechanisms, we used VoIP traffic generated through the IP SLA (Service Level Agreement) protocol. The evaluation criteria we considered include latency, jitter, the MOS (Mean Opinion Score), and packet loss rate. The results of this research are of great significance for network administrators as well as Internet Service Providers (ISPs). They provide valuable insights for IPv6 migration planning within networks, thereby enabling a more efficient and reliable transition to IPv6.
Keywords:	IPv6, Manual IPv6 tunnel, 6rd, GNS3,IP SLA, VoIP, Scalability
Source:	Journal of Theoretical and Applied Information Technology 15^th Decmeber 2023 -- Vol. 101. No. 23-- 2023
Full Text

Title:	MULTI-CLASS CARDIOVASCULAR DISEASES CLASSIFICATION SYSTEM USING DENSE INCEPTION ATTENTION NEURAL NETWORK
Author:	MR. AFROZ PASHA, DR.NAGARAJA S R
Abstract:	In the entire world, cardiovascular diseases (CVD) are the main cause of death compared to other disease. Before experiencing a catastrophic heart failure event such as a stroke, heart attack, or myocardial infarction, people with CVDs may not receive a diagnosis. To address the above problems, To address the above problems, this paper proposes a Dense Inception Attention Neural Network (DIAN-Net) for the Classification of Cardiovascular Diseases using ECG Signals, We add residual blocks, and residual convolutional layer pathways are integrated into the atrous spatial pyramid pooling (ASPP) module and Multi-Scale Context Fusion Block (MSCFB). To fuse convolutional feature maps in encoding layers, the ASPP unit used a learnable set of parameters. An efficient architecture for feature extraction during the encoding step is the ASPP unit. We integrated the AD unit with the benefits of the U-Net network for deep and shallow features. The proposed decoder takes advantage of the multi-scale features from the encoder to predict CVD regions. The aforementioned tests demonstrate that the newly created deep learning models may be very helpful in clinical decision-making and nuclear medicine. 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:	ECG Signal, CNN, Dense Layer, Cardiovascular Diseases.
Source:	Journal of Theoretical and Applied Information Technology 15^th Decmeber 2023 -- Vol. 101. No. 23-- 2023
Full Text

Title:	EXPLORING THE INFLUENCE OF ARTIFICIAL INTELLIGENCE TECHNOLOGY IN MANAGING HUMAN RESOURCE MANAGEMENT
Author:	MYAGMARSUREN OROSOO, NAMJILDAGVA RAASH, KATHARI SANTOSH, DR. CHAMANDEEP KAUR, DR. JEHAD M. ALKHALAF BANI-YOUNIS, MANIKANDAN RENGARAJAN
Abstract:	Numerous areas have conducted in-depth study on artificial intelligence. The world s growing reliance on technology in the context of globalization emphasizes how crucial it is for businesses to be innovative and competitive. The field of Human Resource Management (HRM), has become more important than it has ever been, especially when it comes to hiring workers who can provide an organization with invaluable expertise and skills. With the use of cutting-edge technology, many operations that were formerly completed by hand may now be automated. As such, it is essential to thoroughly examine and assess how technology is affecting the human resource management industry. A theory covering six important areas of HRM was established in an attempt to close the gap between AI and HRM. Human resource planning and strategy, hiring and selecting procedures, skill-development techniques, performance appraisal, pay appraisal, and staff engagement management are some of these domains. The potential use of AI technology is interwoven with these disciplines. This study s main goals were to investigate artificial intelligence s application to human resource management and to learn more about the difficulties that human resources departments encounter. The results of the study showed that AI is crucial to many HR functions, including hiring, data analysis, gathering data, and job fulfilment. This emphasizes how AI is becoming more and more important in improving the efficacy and efficiency of HRM procedures.
Keywords:	Artificial Intelligence; Human Resource Management; Recruitment; Training; Performance Management.
Source:	Journal of Theoretical and Applied Information Technology 15^th Decmeber 2023 -- Vol. 101. No. 23-- 2023
Full Text

Title:	CONTENT BASED VIDEO RETRIEVAL USING LOW-LEVEL FEATURES
Author:	SUBRAMANYAM KUNISETTI, SUBAN RAVICHANDRAN
Abstract:	Over the past ten years, researchers have studied the Content Based Video Retrieval System, based on numerous applications, advancements, and technologies. In video retrieval systems, there is still a requirement for the high-level semantic elements as and the processing of low level materials. As a result, it inspires and motivates a lot of academics to learn more about the content Retrieval and to make more useful and effective while creating system applications. Analysis of video for retrieval of key aspects is regarded as earlier work. In this instance, input videos from YouTube are watched for analysis. After that, foreground segmentation is carried out to find a lot of tiny subset objects since each subset must identify the foreground video class. Given for feature extraction is the split region. In this study, four distinct methods for extracting features—chromatic moment, blur, color variety, and reflection features are taken into consideration. The high dimensionality characteristics are removed from the retrieved features using Principal Component Analysis since they may affect classification accuracy. The feature vectors are taken into account while combining all of the parts. The Nave Bayes classifier is used to complete the classification process. Metrics including accuracy, precision, recall, and F-measure are used to gauge how effectively video retrieval is performed. The predicted model outperforms the current strategy when the proposed model and the in-depth learning approach are compared.
Keywords:	Content Retrieval, Foreground Segmentation, Image Retrieval, Chromatic Moment, Naïve Bayes Classifier.
Source:	Journal of Theoretical and Applied Information Technology 15^th Decmeber 2023 -- Vol. 101. No. 23-- 2023
Full Text

Title:	REVOLUTIONIZING SENTIMENT ANALYSIS IN LITERARY TEXT THE IMMORTALS OF MELUHA THROUGH A HYBRID CNN-RNN ARCHITECTURE AND ADVANCED FEATURE TECHNIQUES
Author:	MANIKANDAN RENGARJAN , DR.BINDU M.R , DR. VIJAYALAKSHMI PONNUSWAMY , DR. GULNAZ FATMA
Abstract:	Sentiment analysis has emerged as a pivotal task in natural language processing, enabling the automated interpretation of emotions and opinions expressed in text. The study presents a novel method for summarizing and sentiment analysis of the literary work "The Immortals of Meluha." A hybrid CNN-RNN architecture is used in the suggested methodology, along with sophisticated feature extraction and selection procedures that makes use of the Artificial Gorilla Troops Optimizer (AGTO) model. The first stage extracts the textual information, followed by preprocessing techniques including tokenization, stemming, and the removal of special characters, stop words, and nulls. Two methods are used in feature extraction: the Term Frequency-Inverse Document Frequency (TF-IDF) methodology, which measures word significance in the dataset, and the Assimilated N-Gram (ANG) method, which gathers contextual data. The most distinctive characteristics are found using the AGTO model, which optimally choose characteristics and improve the effectiveness of classification. The development and execution of a hybridized CNN-RNN architecture, which combines Long-Short Term Memory (LSTM) and Convolutional Neural Networks (CNN), is the basis of the present paper. This design helps with sentiment evaluation and summarizing positions by capturing local as well as international relationships in the textual material efficiently. The results of the experimental assessment show that the CNN-RNN methodology outperforms other approaches, including word2vec+CNN, doc2vec+LR, and one-hot+LR with higher accuracy of 98%. The suggested method s effectiveness is further confirmed by measures such as area under the curve (AUC), precision, recall, and F-measure. The study is unusual because it takes an integrated strategy, combining a hybridized deep learning architecture with sophisticated preprocessing, feature extraction, and selection algorithms. The important contribution of the suggested technique to natural language processing—particularly in the areas of sentiment analysis and summarization—is highlighted in the study s conclusion.
Keywords:	Artificial Gorilla Troops Optimizer, Deep Neural Networks, Meluha, N-Gram Approach, Sentiment Analysis
Source:	Journal of Theoretical and Applied Information Technology 15^th Decmeber 2023 -- Vol. 101. No. 23-- 2023
Full Text

Title:	GENERATION OF AUTOMATED TEXTS AND REPORTS FOR THE CASE OF INFLATION IMPACT ON INDUSTRIES: AN APPROACH BASED ON DEEP LEARNING
Author:	A. AL KARKOURI , M. LAZRAK ,F. GHANIMI , H. EL AMRANI , D. BENAMMI, S. BOUREKKADI
Abstract:	The paper proposes an innovative technique based on deep learning to automate the development of texts and reports addressing the effect of inflation on various economic sectors. Strategic decision-making is dependent on access to trustworthy and timely information, and the effect of inflation is vital for firms, consumers, and government decision-makers. Inflation analysis and report writing are introduced as two of the many obstacles economic analysts encounter and are discussed at length. It then presents deep learning as an effective method for turning data into useful information. The "Materials and Methods" section provides in-depth explanations of dataset construction, data preprocessing, model development, training, and assessment. These vital procedures are required to guarantee the accuracy of the results. The essay highlights how the deep learning technique may boost the precision of economic research while increasing the velocity with which texts and reports can be generated, providing key data to decision-makers in near real-time. The success of this approach is shown by real-world instances of automatically produced reports across many industries. The essay explains how automation is becoming more significant in economic research. It s a glimpse into the future of automated economic research, showing how deep learning is transforming our comprehension and use of inflation-related data. This development has great potential as a useful new resource for economic players and decision-makers by expeditiously disseminating critical information.
Keywords:	Deep Learning, Generation Text And Reports, Price Increase, Inflation On Industries
Source:	Journal of Theoretical and Applied Information Technology 15^th Decmeber 2023 -- Vol. 101. No. 23-- 2023
Full Text

Title:	TECHNICAL IMPLICATIONS OF SMART CONTRACT ARCHITECTURE FOR SHIPBUILDING STAGES AND PAYMENTS
Author:	KYUNGHWAN KIM, SANGSEOP LIM, CHANG-HEE LEE, SEOK-HUN KIM
Abstract:	Smart contracts are a technology that has been applied, or is being considered for application, in a variety of sectors to expand the way we transact based on the trustworthiness of the blockchain, beyond the basic functionality of simple payment methods such as the existing Bitcoin-based blockchain. However, there are limitations to the application of smart contracts in shipbuilding contracts due to the complexity of the transaction steps. Therefore, this study examined the applicability of blockchain-based smart contracts to shipbuilding contracts by comparing the legal precedents and jurisprudence of English law, which is most often used as the governing law for shipbuilding contracts, and Korean law, which is most often used for shipbuilding contracts. Through this, we identified major legal issues from the collateralization and prevention perspectives of legal stability, and proposed improvements by applying the conceptual level architecture and algorithms of smart contracts in shipbuilding contracts to the payment conditions. While this study was not able to program the algorithm to apply to all shipbuilding contracts, it is important to note that the study examines the legal issues that can be expected from a legal perspective.
Keywords:	Block Chain, Smart Contract, Shipbuilding Contract, Contract Algorithm, Shipbuilding Payment
Source:	Journal of Theoretical and Applied Information Technology 15^th Decmeber 2023 -- Vol. 101. No. 23-- 2023
Full Text

Title:	SUPERVISION TRANSFORMATION IN GOVERNMENT INSTITUTIONS
Author:	DINNA MAYASARI, ARRY AKHMAD ARMAN
Abstract:	The supervision process in government starts from the planning, implementation, monitoring, and evaluation stages. Supervision is measured by how compliant an organization is in complying with statutory regulations. Currently checking data is only limited to data retrieval. Accelerated regulatory changes in government can create legal issues if organizations are unable to meet regulatory compliance. Limited Human Resources (HR) in analyzing the many regulations and limited inspection time result in less-than-optimal quality of supervision. Supervisors oversee complex problems and often have difficulty obtaining data and information to analyze to produce fast, efficient, and effective decisions. The current monitoring system runs on a computer-based system, but there are still many opportunities for improvement. Some opportunities that can be implemented include (1) end-to-end integration of all processes, (2) use of technology to check data validity, (3) reduction of approval processes, (4) early warning systems, and so on. The supervisory process of the Financial Services Authority (OJK, Otoritas Jasa Keuangan) has begun and the concept of implementing Regulatory Technology (RegTech) and Supervisory Technology (SupTech) has begun in an integrated manner on the supervisory and regulatory industry side. The results of the literature study have not answered the standard architectural solutions that can be used as a reference for monitoring solutions in government agencies. In this research, a similar solution is designed for monitoring processes in government agencies. In this research, a solution architecture artifact was designed for a monitoring system in government institutions. The chosen methodology is Design Science Research Methodology (DSRM). The designed solution is represented by adopting the 4-layer divisions used in Enterprise Architecture (EA), namely the business layer, data layer, application layer, and technology layer. Testing takes the form of measuring compliance with requirements and simulating business processes to predict the impact on the performance of the system being designed.
Keywords:	Architecture, RegTech, SupTech, Technology, Government
Source:	Journal of Theoretical and Applied Information Technology 15^th Decmeber 2023 -- Vol. 101. No. 23-- 2023
Full Text

Title:	VIRTUALIZATION IN CLOUD SECURITY
Author:	JAWHARA BOODAI, AMINA ALQAHTANI, MOUNIR FRIKHA
Abstract:	Cloud computing has become ubiquitous, providing convenient on-demand access to computing resources. However, security remains a major concern, especially with the added complexity of Virtualization. This paper systematically reviews research on virtualization security in cloud environments. We surveyed academic literature from 2010-2023 to summarize the latest techniques and algorithms to secure virtualized cloud infrastructure and prevent attacks. Common methods include hypervisor hardening, micro-segmentation, virtual network encryption, and virtual machine introspection. Research trends point towards increasingly advanced techniques like homo-morphic encryption and confidential computing, enabling secure, privacy-preserving computation on encrypted data. More work is still needed to balance performance and scalability with security. This review provides an overview of the state-of-the-art securing virtualized cloud environments and identifies open challenges for future research.
Keywords:	Cloud Security, Virtualization, Hypervisor, Micro Segmentation, Encryption.
Source:	Journal of Theoretical and Applied Information Technology 15^th Decmeber 2023 -- Vol. 101. No. 23-- 2023
Full Text

Title:	AUTOMATING DATA WAREHOUSE DESIGN WITH MDA APPROACH USING NOSQL AND RELATIONAL SYSTEMS
Author:	LAMYA OUKHOUYA, ANASS EL HADDADI,BRAHIM ER-RAHA, ASMA SBAI
Abstract:	Data warehouses and OLAP systems provide methods and tools for analyzing data from enterprise information systems. Unfortunately, relational data warehouses are unable to store and analyze data with the 3V characteristics of Big Data: volume, variety, and velocity. To address this, NoSQL systems are introduced in addition to RDBMS, offering scalability to data warehouses to effectively adapt to the volume and variety of collected data. However, integrating these two systems in the same architecture in Big Analytics processes is complex, both in terms of data modeling and data processing. In this regard, several approaches have been proposed to alleviate this complexity. However, several points, which relate to integrated modeling abstractions, adapting the conceptual model with various NoSQL and relational systems, or automating the design process, remain unexamined. In this article, our approach Accounts for all these limitations through a model-driven architecture approach (MDA). This approach proposes a design with three levels of abstraction: conceptual, logical, and physical. The conceptual level is presented by a multidimensional model. The logical level is described by a generic model for all NoSQL and relational families, and the physical level is described by three models related to the implementation; MySQL DBMS for relational systems, and Cassandra DBMS and MongoDB DBMS for NoSQL systems. Moreover, the entire design process is automated through a set of implemented transformation rules rom the conceptual model to source code extraction, thereby facilitating the design task for developers. Furthermore, we conducted a qualitative evaluation compared to other methodologies, revealing that our approach excels in using a generic logical model that can be adapted at the physical level to five types of NoSQL systems and relational systems. Additionally, the automation of the transition from a conceptual model to source code extraction offers a notable advantage in simplifying the migration between concepts.
Keywords:	Data Warehouse, Relational Systems, Nosql Systems, Hybrid Architecture, Model-Driven Architecture.
Source:	Journal of Theoretical and Applied Information Technology 15^th Decmeber 2023 -- Vol. 101. No. 23-- 2023
Full Text

Title:	OPTIMIZING RESIDENTIAL ENERGY CONSUMPTION THROUGH CAE-LSTM
Author:	BADARI NARAYANA PALETY, DR. C. MAHALAKSHMI, DR. P. NAGASEKHAR REDDY
Abstract:	In the modern world, limiting domestic energy consumption is a crucial task. Finding effective ways to use energy in our homes is crucial as energy demands rise and concerns about climate change increase. As a consequence of this optimisation, homeowners save a significant amount of money while simultaneously reducing their carbon footprint. In order to solve this problem, this study combines Long Short-Term Memory (LSTM) and Convolutional Autoencoder (CAE) neural networks. With the use of deep learning and sequence modelling, the suggested CAE-LSTM framework can intelligently control and lower energy consumption in residential structures. Time-series energy data are subjected to feature extraction and dimensionality reduction utilising the CAE, enabling the discovery of hidden patterns and abnormalities. Contrarily, the LSTM network captures the temporal relationships in energy use patterns, enabling precise forecasts and proactive energy management. This study makes utilisation of a large dataset of information on the energy consumption of households.On the basis of this dataset, the CAE-LSTM model is trained to learn complicated correlations between many variables, including weather, occupancy patterns, and appliance utilise, and how these variables affect the consumption of electricity. The experimental findings show that the CAE-LSTM model is capable of learning and adapting to changing energy consumption patterns, leading to significant energy savings and increased sustainability. A more energy-efficient and ecologically friendly future could result from the widespread adoption of smart energy management systems in residential settings as a consequence of this investigation.
Keywords:	Convolutional Auto Encoder Long Short-Term Memory (CAE-LSTM); Convolutional Autoencoder (CAE); Long Short-Term Memory (LSTM); Energy management; Residential Energy Consumption
Source:	Journal of Theoretical and Applied Information Technology 15^th Decmeber 2023 -- Vol. 101. No. 23-- 2023
Full Text

Title:	TIME SERIES CLASSIFICATION FOR FINANCIAL STATEMENT FRAUD DETECTION USING RECURRENT NEURAL NETWORKS BASED APPROACHES
Author:	ITO WASITO, FEBRYANTI SIMON, PRITA KARINA DIANDRA
Abstract:	Financial statements fraud, considered as untruthful behavior for gaining financial benefits, has recently become a widespread issue in companies and organizations. With the advancement of artificial intelligence, deep-learning-based approaches can be used intelligently to detect fraudulent transactions by analyzing a large number of financial statements data. This paper proposes to formulate financial statements fraud detection into time series classification (TSC) problems using Recurrent Neural Networks (RNN) based approaches which include Simple RNN, Long Short-Term Memory (LSTM) and Gated Recurrent Unit (GRU). The implementation of the proposed approaches will be demonstrated on financial statements data collected from Accounting and Auditing Enforcement Releases (AAERs), which are federal materials issued by US Security and Exchange Commission (SEC). The objective will be achieved through two steps. First, the Time Series Classification (TSC) and RNN based approaches will be reviewed. Then, the experimental settings of RNN based approaches comparisons on specified data sets will be introduced. As it is a TSC problem, therefore the evaluation of the RNN based model performances will be determined based on percentage of accuracy and loss-accuracy curves as visualization tool. This study has two contributions: first, this research can provide insight to corporate management and investors in detecting fraud that happens in the company, second contribution for the academic purpose, this research proposes alternative methods in detection of fraud. The results show that on various hidden unit number, GRU architecture model outperforms Simple RNN and LSTM. The best performance is achieved by GRU model at 14 hidden unit number which produces more than 99% in training accuracy and more than 91% in test accuracy. Overall, Simple RNN to be moderate model and LSTM is the worst model.
Keywords:	Time Series Classification, Fraud, Financial Statement, Recurrent Neural Networks, Deep Learning
Source:	Journal of Theoretical and Applied Information Technology 15^th Decmeber 2023 -- Vol. 101. No. 23-- 2023
Full Text

Title:	THE IMPACT OF OBJECTIVE FUNCTIONS ON TASK SCHEDULING IN CLOUD COMPUTING ENVIRONMENT
Author:	NORA OMRAN ALKAAM ,ABU BAKAR MD. SULTAN , MASNIDA HUSSIN, KHAIRONI YATIM SHARIF
Abstract:	Cloud computing, which has grown in popularity in recent years, allows users to use computational resources remotely over the Internet. Cloud computing must be able to meet all user demands for high performance and efficient service quality (QoS). As a result, in order to meet these requests in a timely manner, an effective task scheduling mechanism must be created. The aim of this study is to explore the current landscape of task scheduling problems, laying out the challenges of task scheduling where objective functions issues are involved. We used a systematic literature review strategy to locate and review many significant journal and conference papers on four major online electronic databases (ScienceDirect, IEEE Explore, Springer, Wiley online library) that addressed our three predefined study questions. The first stage was to define inclusion and exclusion criteria before extracting data from the selected publications and deriving replies to our inquiries. Finally, (75) publications were chosen. We identified (70) publications on task scheduling describing (58) investigations on objective functions published between 2018 and mid-2022. Findings show a trend across work scheduling algorithms to choose diverse objective functions. These algorithms often optimize for time efficiency, cost-effectiveness, and resource use. In contrast, some algorithms specialize in a single objective function. This difference in methodology suggests that task scheduling performance depends on the objective function. The algorithm s effectiveness and adaptability in cloud-based job scheduling depend on these objectives careful selection.
Keywords:	Cloud Scheduling, Multi-Objective Functions, Single-Objective Function, Task Scheduling, Cloud Computing.
Source:	Journal of Theoretical and Applied Information Technology 15^th Decmeber 2023 -- Vol. 101. No. 23-- 2023
Full Text

Title:	MULTITASK LEARNING FOR GENDER IDENTIFICATION AND AGE GROUP BASED ON THE MANDIBLE ON PANORAMIC RADIOGRAPHS
Author:	NUR NAFIIYAH, CHASTINE FATICHAH, DARLIS HERUMURTI, EHA RENWI ASTUTI, RAMADHAN HARDANI PUTRA
Abstract:	Forensic odontology is commonly applied for victim identification using comparing antemortem and postmortem dental radiographs. However, in cases where a victim s teeth are incomplete or missing, the mandible bone can also be used as a robust alternative for victim identification. Gender identification and age estimation are two tasks to assist in victim identification. For multiple related tasks, the multitask learning (MTL) approach has been proven to enhance generalization performance by concurrently learning the multiple related tasks and leveraging useful information across the tasks. Therefore, in this study, we propose an MTL approach for gender identification and age group based on the mandible. We propose a model, namely the mandible radiographs MTL model, that takes panoramic radiographs of the mandible as input. We built a dataset, namely the mandible radiographs dataset comprising 120 patients panoramic radiographs of the mandible collected from Universitas Airlangga Dental Hospital, Surabaya, Indonesia, then augmented to 600 images. The experimental results show that the augmented mandible radiographs MTL model achieved the best performance for gender identification with a mean accuracy of 99.7% and an age group of 99.5%. Our research proposal is more practical because 1 model directly produces two outputs (gender and estimated age), so it is time efficient in creating models or testing.
Keywords:	Multitask Learning, Mandibular Panoramic Radiographs, Dental Panoramic Radiographs, Gender Identification, Age Group.
Source:	Journal of Theoretical and Applied Information Technology 15^th Decmeber 2023 -- Vol. 101. No. 23-- 2023
Full Text

Title:	APPLICATION OF FUZZY LOGIC IN WEAVING PROCESS: A SYSTEMATIC LITERATURE REVIEW
Author:	R. MESSNAOUI, M. EL BAKKALI, A. SOULHI, O. CHERKAOUI
Abstract:	The process of weaving continues to be one of the most intricate chains of transformation within the textile industry. This complexity arises from the diverse range of structures, the multiple stages involved, the intricate machinery utilized, the utilization of various materials, and the combination of both creativity and precision. Consequently, there exists a necessity for tools that can enhance efficiency, flexibility, and decision-making within this field. This review of existing literature aims to provide pertinent information regarding the utilization of logic in the realm of weaving. In this research study, a systematic literature review methodology was employed to examine the application of fuzzy logic within the weaving process. Data for this study was collected from reputable databases, including ScienceDirect, IEEE Xplore, Textile Research Journal, and Google Scholar. To select relevant articles for this study, the Prisma framework was utilized, resulting in the inclusion of solely journal articles for the literature review. A comprehensive framework was developed to elucidate the impact of employing fuzzy logic, the approach presented in this framework provides a comprehensive and highly effective method for tackling the complex challenges associated with ambiguity, modification, and subtlety that are frequently observed in the intricate and intricate process of weaving. The findings of various studies explored aspects such as the properties of warp and weft fabrics, the performance of weaving machines, the organizational performance of weaving companies, and occupational health and safety concerns. While these studies have provided valuable solutions, they unfortunately remain insufficient in the face of the weaving process, which persists as a complex field characterized by uncertainties, variations, and intricacies that are inherent to the practice. This uniqueness of each woven product fosters an environment conducive to experimentation and further research.
Keywords:	Fuzzy logic, Warp, and weft fabrics, weaving loom, weaving company, working conditions, Weaving productivity.
Source:	Journal of Theoretical and Applied Information Technology 15^th Decmeber 2023 -- Vol. 101. No. 23-- 2023
Full Text