		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 in case of review papers).

Journal of Theoretical and Applied Information Technology
May 2024 | Vol. 102 No.9

Title:	REAL TIME OF CYBERSECURITY RISKS DETECTION APPROACH FOR BLOCKCHAIN BASED PERSONAL MEDICAL DEVICES
Author:	TAYSEER ALKHDOUR, MOHAMMED AMIN ALMAIAH, AITIZAZ ALI, ROMEL AL-ALI, ABDALWALI LUTFI, MOHAMMAD MANSOUR AL-KHASAWNEH, MAHMAOD ALRAWAD, TING TIN TIN
Abstract:	Recent methods for storing and disseminating medical data limit user access to electronic health records (EHR). It lowers care providers’ access to vital information and ultimately creates a barrier to transitioning from traditional healthcare to a digital healthcare system. Numerous cloud-based systems are used for digital healthcare data allocation, but such an approach relies on third-party software such as the cloud. With the advent of industry 4.0 technologies, blockchain enables a decentralized and trustless environment by removing centralized authority. Existing models mainly utilize blockchain as a data storage tool rather than a security platform. Biomedical and monitoring devices generate massive amounts of data, and the existing approach overloads the blockchain with IoT data. This research proposes blockchain as a unique method for securing patient-related data access and integrating homomorphic encryption with an end-to- end privacy-protecting system. In this research, we propose a blockchain-based architecture for identifying security threats in personal medical devices to address the existing issues related to healthcare devices. The proposed framework uses certificate authority to assign an access control token in order to access a particular session. A certificate authority is the nodes based on the reputation within the blockchain network elected through consensus protocol. Proposed framework uses dual certificate authorities, which leads to more reliability and security if one certificate authority is down. Moreover, the existing algorithm overburden the medical devices which are resource constraint such as power oriented and such approaches leads to storage and communication cost overhead. By minimizing latency, security, and data ownership, the proposed framework outperforms the existing centralized system, by comparing the framework and evaluating its performance with the benchmark models.
Keywords:	Cybersecurity, Cyber-Risk Assessment, Authentication, Blockchain, Smart contracts, Latency, Optimization, Security, Health-care.
Source:	Journal of Theoretical and Applied Information Technology 31^st May 2024 -- Vol. 102. No. 10-- 2024
Full Text

Title:	RGB IMAGE ENCRYPTION ALGORITHM USING 6D HYPERCHAOTIC SYSTEM AND FIBONACCI Q-MATRIX
Author:	HUSSEIN YOUNIS, MUJAHED ELEYAT
Abstract:	The advancement of ICT and the widespread use of the Internet enables users to store many types of data via the Internet, and thus these data may be vulnerable to hacking or illegal access. As a result, certain mechanisms must be designed to guarantee that these data are secured. Encryption is one of the most significant methods of image protection as it's based on generating a different image in terms of content from the original image, making it difficult to identify or retrieve the original image only through a key. In this work, an RGB Image Encryption Algorithm Using a 6D Hyperchaotic System and Fibonacci Q-matrix is presented. Confusion and diffusion are used to generate the encrypted image to achieve a high-security level. The proposed encryption algorithm was tested against different types of noise and attacks. Moreover, we combined the proposed algorithm against other algorithms in terms of entropy. The result shows that the proposed algorithm outperformed the existing algorithms. Also, the result shows that the speed up in the execution time of the proposed algorithm can be enhanced up to 2.23x faster by employing parallel programming.
Keywords:	RGB Image Encryption, 6D Hyperchaotic Chen System, Fibonacci Q-Matrix, Parallelism, Multiprocessing.
Source:	Journal of Theoretical and Applied Information Technology 31^st May 2024 -- Vol. 102. No. 10-- 2024
Full Text

Title:	ARABIC/ENGLISH MACHINE-PRINTED AND HANDWRITTEN TEXT IDENTIFICATION IN DOCUMENT IMAGES USING IMPRINT TEXTURE AND CNN
Author:	AHMAD A. ALZAHRANI
Abstract:	The identification of language and writing styles, referred to as script identification, is crucial for automated document images analysis. In Arabic-speaking countries, documents often contain machine-printed and handwritten text in both Arabic and English, which poses a challenge for document image digitization and OCR systems. This paper proposes an image processing with a deep learning-based system that can identify the script type (Arabic or Latin) and its nature (printed machine or handwritten) in document images. Firstly, the system produces an imprint image of the text as input to enhance accuracy. Then using a convolutional neural network (CNN) architecture for feature extraction and classification. The system is trained and evaluated based on benchmark datasets such as the Khatt dataset, the IAM Handwriting Database, the Arabic Sentiment Twitter Corpus dataset, and the LRDE Document Binarization Dataset. The results show that the proposed method significantly improves the identification of text type and style compared to the state-of-the-art techniques.
Keywords:	OCR, Text Identification, Document Images, Arabic Language, CNN.
Source:	Journal of Theoretical and Applied Information Technology 31^st May 2024 -- Vol. 102. No. 10-- 2024
Full Text

Title:	EVALUATION OF A PARKING FINDER APPLICATION USING THE USER CENTERED DESIGN (UCD) METHOD
Author:	OKTO SARAGIH, TANTY OKTAVIA
Abstract:	Parking problems in big cities in Indonesia, such as Jakarta, are a concern because they are not only related to inefficient use of fuel and traffic congestion, but also safe parking. Vehicle users still use illegal parking services because the available parking lots are considered inadequate and parking management is inefficient. To overcome this problem, a smart parking system is needed that is effective in finding available parking lots. Currently, there are parking search applications such as Parkiran.id and Parkee that have the main features of finding the nearest parking location, parking reservations, and making parking payments. However, both applications still have shortcomings that provide a poor experience for users, such as the On Street Parking System feature that is often problematic in the Parkiran.id application and the membersip feature that does not work in the Parkee application, and also the digital payment system that is not effective in both applications. Therefore, the User Centered Design (UCD) method is needed to determine user needs for parking search applications. The user experience (UX) evaluation of this parking application will be carried out by involving users, which is done using two methods, namely, the System Usability Scale (SUS) and the User Experience Questionnare (UEQ). Based on the research conducted, the parkee application is superior to the parkiran.id application, and the two applications both received input regarding the lot towards the parking point because the current parking search application only provides directions to the location of the parking lot connected to google maps.
Keywords:	User Experience, User Interface, Parking System, System Usability Scale (SUS), User Experience Questionnare (UEQ)
Source:	Journal of Theoretical and Applied Information Technology 31^st May 2024 -- Vol. 102. No. 10-- 2024
Full Text

Title:	ENHANCING NETWORK INTRUSION DETECTION AND CLASSIFICATION BY USING HYBRID MACHINE LEARNING APPROACHES
Author:	WASEEM AKRAM, ABID IRSHAD KHAN , HINNA HAFEEZ , MUHAMMAD WASEEM IQBAL, NOR ZAIRAH AB RAHIM, YASIR MAHMOOD, MUHAMMAD AAMIR
Abstract:	The present era is the modern technology evolving era for cybersecurity. It boons a dynamic battlefield for cyber security concerns for security experts. Network intrusions have become a major concern in cyberspace for compromising security. Traditional methods like manual rules, blacklists, and whitelists are insufficient for detecting modern intrusions. While machine learning approaches for intrusion detection have emerged, many suffer from low accuracy. However, recent advances in machine learning algorithms show promise for improving intrusion detection and classification. To address the limitations of current methods, this work proposes a hybrid machine learning approach for intrusion detection and classification. The approach utilizes seven classifiers including decision tree, random forest, naïve Bayes, ADA, XGB, KNN, and logistic regression. The model is evaluated on the CICIDS2017 dataset using training and testing splits. The classifiers achieve accuracy rates of 0.99 for decision tree, 0.96 for random forest, 0.85 for naïve Bayes, 0.97 for ADA, 0.96 for XGB, 0.98 for KNN, and 0.91 for logistic regression. The decision tree classifier demonstrates the highest accuracy of 0.99, owing to its effective parametric function evaluation and ability to minimize misclassification errors. The proposed hybrid approach aims to advance network intrusion detection and classification capabilities beyond current techniques.
Keywords:	Intrusion Detection, Machine Learning, Decision Tree, KNN, Random Forest, Naïve Bayes, ADA, XGB
Source:	Journal of Theoretical and Applied Information Technology 31^st May 2024 -- Vol. 102. No. 10-- 2024
Full Text

Title:	A MACHINE LEARNING APPROACH IN COMMUNICATION 5G-6G NETWORK
Author:	SWATI LAKSHMI BOPPANA, ANNAPURNA GUMMADI, YALLAPRAGADA RAVI RAJU, DR. SATISH THATAVARTI, RAMU KUCHIPUDI5, TENALI ANUSHA, RALLABANDI CH S N P SAIRAM
Abstract:	Applications in the fields of entertainment, commerce, health, and public safety rely heavily on wireless communication technologies. These technologies are constantly improving with each new generation, and the latest example of this is the widespread implementation of 5G wireless networks. Industry and academia are already planning the next generation of wireless technologies, 6G, which will be an improvement over 5G. When it comes to 6G systems, one of the most important things is that these wireless networks employ AI and ML. There will be some kind of artificial intelligence or machine learning used in every part of a wireless system that we know about from our experience with wireless technologies up to 5G, including the physical, network, and application levels. A current overview of concepts for future wireless networks, including 6G, and the relevance of ML approaches in these systems is presented in this overview article. Specifically, we set out a 6G conceptual model and demonstrate how ML approaches are utilized and contribute to each layer of the model. With wireless communication systems in mind, we take a look back at many ML methods, both old and new, including supervised and unsupervised learning, RL, DL, and FL. At the end of the article, we touch on some potential future uses and difficulties with 6G network ML and AI research.
Keywords:	Fifth generation (5G), sixth generation (6G), artificial intelligence (AI), machine learning (ML), deep learning (DL), reinforcement learning (RL), federated learning (FL).
Source:	Journal of Theoretical and Applied Information Technology 31^st May 2024 -- Vol. 102. No. 10-- 2024
Full Text

Title:	ARTIFICIAL INTELLIGENCE FRAMEWORK FOR MULTI-CLASS SUGARCANE LEAF DISEASES CLASSIFICATION USING DEEP LEARNING ALGORITHMS
Author:	A.VIVEKREDDY, R.THIRUVENGATANADHAN, M.SRINIVAS, P. DHANALAKSHMI
Abstract:	Sugarcane is a crucial crop in the global agriculture industry, contributing significantly to the production of sugar, ethanol, and other by-products. However, the prevalence of various diseases can severely affect sugarcane yield and quality, making timely and accurate disease detection imperative. Current methods for sugarcane disease detection predominantly rely on traditional image classification models, which often lack the required precision and efficiency, leading to delayed interventions and compromised crop health. To address these limitations, this paper introduces an efficient and novel comparison of Efficient deep learning models that leverages the strengths of multiple deep learning classifiers, including Alex net, ResNet18, VGG19, and Densenet201, for enhanced sugarcane disease prediction. These diseases include Red Rot, Red Rust, mosaic, and yellow leaf disease. In our approach, individual classifiers are initially employed to classify sugarcane images, followed by the identification and comparison of the best-performing classifiers. In this paper, we have used 1990 sugarcane leaf images for the classification of leaf diseases into normal, red rust, red rot, and bacterial blight. or four class classifications, VGG19 had the greatest accuracy of 98.82%, precision of 96.77%, and sensitivity of 96.33%. The implications of this work are profound, as the proposed model significantly outperforms existing methods in terms of efficiency, accuracy, and timeliness. This advancement holds the potential to revolutionize sugarcane disease detection, ultimately contributing to better crop management, improved yields, and enhanced profitability for farmers and stakeholders in the sugarcane industry. The integration of efficient deep learning models in agriculture paves the way for more informed decision-making processes, ultimately safeguarding the livelihoods of those dependent on sugarcane farming.
Keywords:	CNN, Mosaic, Red Rot, Yellow, Red Rust
Source:	Journal of Theoretical and Applied Information Technology 31^st May 2024 -- Vol. 102. No. 10-- 2024
Full Text

Title:	IMPLEMENTATION OF OBFS USING FEATURE EXTRACTION AND INFORMATION GAIN TECHNIQUES FOR SKIN DISEASE CLASSIFICATION
Author:	VEMPATI KRISHNA, DR.Y.RAMAMOHAN REDDY, DR. D. MARUTHI KUMAR, DR.G.HEMALATHA, DR. Y. PRAVEEN KUMAR, M. KHALEEL ULLAH KHAN, RAVINDRA CHANGALA
Abstract:	With the growing prevalence of skin diseases and the ever-increasing potential of computational diagnostics, this study delves into the exploration and comparison of three diverse models—Optimized Biomarker Feature Selection (OBFS), Convolutional Neural Networks (CNN), and PCA-based Classification—for skin disease classification. Utilizing the "DermNet Skin Disease Dataset" as our experimental ground, we evaluated the models on parameters like complexity, interpretability, computational efficiency, and adaptability to new data. The OBFS model, which uniquely combines feature extraction with information gain techniques, displayed a balanced performance, merging interpretability with decent computational demands. The results and insights gleaned from our investigation offer a foundational framework for researchers and practitioners in dermatology, emphasizing the potential and limitations of computational methods in skin disease classification.
Keywords:	Skin disease classification, Optimized Biomarker Feature Selection (OBFS), Convolutional Neural Networks (CNN), PCA-based Classification, Computational diagnostics.
Source:	Journal of Theoretical and Applied Information Technology 31^st May 2024 -- Vol. 102. No. 10-- 2024
Full Text

Title:	INTRUSION DETECTION SYSTEM FOR CYBER SECURITY IN SMART AGRICULTURE WITH ABCIS TECHNIQUES
Author:	NASREEN SULTANA QUADRI, DR. YASMEEN, K DURGA CHARAN, K SURESH BABU, DR. SHAHANA TANVEER, K. SHYAM SUNDER REDDY, DR.M. KIRAN KUMAR
Abstract:	In this research, we examine and evaluate intrusion detection systems for cyber security in Agriculture 4.0. In particular, we outline the assessment criteria and cyber security risks that are utilised to assess an intrusion detection system's effectiveness for Agriculture 4.0. Then, we assess intrusion detection systems in light of cutting-edge technological developments, such as cloud computing, fog/edge computing, network virtualization, Internet of Things, autonomous tractors, drones, industrial agriculture, and smart grids. We offer a thorough classification of intrusion detection systems in each developing technology, based on the machine learning approach utilised. In addition, we provide public datasets and the frameworks used for implementation that were used to assess intrusion detection systems' performance for Agriculture 4.0. Lastly, we discuss the obstacles and potential lines of inquiry for future studies in intrusion detection for cyber security in Agriculture 4.0. Based on several technical paradigms, a new industrial revolution is underway. "Industry 4.0" (I4.0) is a concise way to communicate the desire to promote and direct this phenomena. Projects falling under this umbrella term are united by the belief that numerous critical technologies supporting Big Data Analytics and Cyber-Physical Systems are merging to form a new, highly automated, distributed, and dynamic production network. To ensure that this process proceeds smoothly and on schedule, new laws and cultural norms must be put in place. In this paper, we exclusively address the technological side, emphasising the exceptional I4.0 complexity that has been documented in the scientific literature.
Keywords:	ABCIS, IoT, blockchain, cyber intrusion detection, cloud computing, AI, SDN.
Source:	Journal of Theoretical and Applied Information Technology 31^st May 2024 -- Vol. 102. No. 10-- 2024
Full Text

Title:	EVOLVING EDUCATIONAL TECHNOLOGIES: A BIBLIOMETRIC STUDY ON AI IN INTELLIGENT TUTORING SYSTEMS
Author:	CANUJA C. S, JOBIN JOSE, NANCY J FERNANDEZ, BINDU JOSEPH, LIZ KURIAKOSE, NIDHU NEENA VARGHESE
Abstract:	A metamorphic advancement has been made in the field of educational technology with the integration of Artificial Intelligence (AI) and Intelligent Tutoring Systems (ITS). This integration equips a tailored and robust learning gest. This paper uses an inclusive bibliometric analysis to examine the convergence of ITS and AI, focusing on the Scopus database for gathering bibliographic data. The procedure is designed over the PRISMA flow chart, which accurately deals with the inclusion and exclusion criteria for the literature review. This study utilizes advanced bibliometric tools such as Biblioshiny, VOSviewer, and CiteSpace, providing a multifaceted view of the field's development and impact. The analysis yields significant insights into various aspects: the annual scientific production in the field, identifying the most influential authors, and discerning the most relevant sources. Furthermore, it highlights the most globally cited documents, pinpoints the most productive country in this research area, and delves into the trending topics. Additionally, the study incorporates a thematic map, factorial analysis, bibliographic coupling of documents, and the co-occurrence of all keywords, providing a multidimensional view of the field. The analysis categorizes keywords with prominent citation bursts, studies the co-citation of cited authors, and engages network visualization techniques for cited journals. A timeline visualization of international collaborations suggests further acumens into the global subtleties of the field. The paper concludes by finding appropriate research gaps and discussing the application of these findings, thus contributing to further research in the field of educational technology.
Keywords:	Intelligent Tutoring System, Artificial Intelligence, Bibliometric Analysis, Biblioshiny, VOSviewer, Citespace
Source:	Journal of Theoretical and Applied Information Technology 31^st May 2024 -- Vol. 102. No. 10-- 2024
Full Text

Title:	DISTRIBUTED LEDGER TECHNOLOGY ON ENHANCED PEGASIS ALGORITHM FOR MOBILITY MANAGEMENT IN WSN
Author:	CH GANGADHAR, HABIBULLA MOHAMMAD, K. PHANI RAMA KRISHNA, RIAZUDDIN MOHAMMED
Abstract:	Today, in order to monitor and observe many factors in any environment wireless sensor technology is being widely used. Sensor node is the major component of WSN which lacks in its limited energy resources and lifetime duration. The deployment of sensors takes place in a hostile environment that is typically unavailable for the replacement of worn-out batteries. As a result the sensors have to use their original energy to do those tasks. Due to this restriction, energy efficiency is a crucial WSN characteristic. To get over these constraints Distributed Ledger Technology (DLT) and Enhanced PEGASIS (Power-Efficient Gathering in Sensor Information Systems) Algorithm can be integrated to enhance mobility management in Wireless Sensor Networks (WSNs). DLT, commonly associated with block chain technology, offers a decentralized and immutable ledger for recording transactions across a network of nodes. While PEGASIS is an energy-efficient data gathering algorithm that organizes sensor nodes into a chain topology and employs a greedy algorithm to minimize energy consumption during data transmission. For enhancing mobility management, energy efficiency, security, and resilience in Wireless Sensor,an overall, integrating Distributed Ledger Technology with the Enhanced PEGASIS Algorithm was introduced. The results revealed that the network lifetime has improved using DLT-EPEGASIS when compared to the previous protocols like LEACH and PEGASIS
Keywords:	WSN, Energy Efficiency, EPEGASIS, LEACH
Source:	Journal of Theoretical and Applied Information Technology 31^st May 2024 -- Vol. 102. No. 10-- 2024
Full Text

Title:	ADVANCEMENTS IN DYNAMIC TOPIC MODELING: A COMPARATIVE ANALYSIS OF LDA, DTM, GIBBSLDA++, HDP AND PROPOSED HYBRID MODEL HDP WITH CT-DTM FOR REAL-TIME AND EVOLVING TEXTUAL DATA
Author:	C.B.PAVITHRA, DR.J.SAVITHA
Abstract:	This research presents a comprehensive analysis of dynamic topic modeling approaches applied to the intricate task of modeling real-time and evolving textual data. It investigates five distinct methodologies, including Latent Dirichlet Allocation (LDA), Dynamic Topic Modeling (DTM), Latent Dirichlet Allocation with Gibbs Sampling (GibbsLDA++), the Hierarchical Dirichlet Process (HDP), and our innovative Hybrid approach combining Hierarchical Dirichlet Process (HDP) with Continuous-Time Dynamic Topic Modeling (CT-DTM). The primary objective of this study is to evaluate the effectiveness of these methods in capturing, tracking, and adapting to the ever-changing landscape of topics and trends within a wide range of textual datasets, spanning social media conversations, news articles, scientific publications, and beyond. The goal is to evaluate their efficacy in capturing the evolving themes within a corpus of research papers, providing insights into the strengths, limitations, and potential use cases for each model. The research aims to gain insights into the unique strengths and limitations of each technique, examining their interpretability, computational efficiency, and adaptability to evolving data distributions. Furthermore, the research explores the potential enhancements achieved by hybridizing HDP with CT-DTM, offering an approach that combines structured topic modeling with continuous-time modeling. This investigation is particularly timely, given the dynamic nature of contemporary (modern) data sources and the critical need for models that can flexibly adapt to emerging trends and shifting textual patterns. The findings of this research provide valuable insights into the suitability of LDA, DTM, GibbsLDA++, HDP, and the groundbreaking Hybrid HDP and CT-DTM approach for dynamic topic modeling in real-time and evolving textual data.
Keywords:	Dynamic Topic Modeling, Latent Dirichlet Allocation (LDA), Dynamic Topic Modeling (DTM), Gibbslda++, Hierarchical Dirichlet Process (HDP)
Source:	Journal of Theoretical and Applied Information Technology 31^st May 2024 -- Vol. 102. No. 10-- 2024
Full Text

Title:	A NOISE-ROBUST DESCRIPTOR: APPLICATIONS TO FACE RECOGNITION
Author:	ABDELLATIF DAHMOUNI, ABDELKAHER AIT ABDELOUAHAD, HASSAN SILKAN
Abstract:	In the field of computer vision research, identifying the distinctive features of image objects remains a persistent challenge, particularly in face recognition. Although most face recognition techniques excel under ideal conditions, they are limited to the various changes that affect the accuracy of face feature extraction, such as variations of lighting, facial expressions, poses, occlusions, and noise. To best circumvent this problem, we propose a new variant of the Local Binary Pattern (LBP) called the Robust Electric Virtual Binary Pattern (REVBP). The main purpose of the REVBP is to improve the noise performance of the Electric Virtual Binary Pattern (EVBP) descriptor using the localized noise coding method. In order to assess the effectiveness of the proposed approach, two processes are used. The first is a comparative study between REVBP and EVBP in terms of recognition rate and processing time. The second involves combining REVBP with various machine learning algorithms, including variants of the Support Vector Machine (SVM). Extensive experiments have shown that REVBP performs better than the original EVBP. They also showed that using learning classifiers provided significant improvements in terms of recognition accuracy, outperforming several alternative methods.
Keywords:	Enhanced Local Binary Pattern Histogram (eLBPH); Robust Electric Virtual Binary Pattern (REVBP); Electric Virtual Binary Pattern (EVBP); Support Vector Machine (SVM).
Source:	Journal of Theoretical and Applied Information Technology 31^st May 2024 -- Vol. 102. No. 10-- 2024
Full Text

Title:	TOWARDS A UNIFIED FORMALIZATION OF OPACITY PROPERTIES IN DISCRETE AND REAL-TIME SYSTEMS
Author:	HAFEDH MAHMOUD ZAYANI, IKHLASS AMMAR, MOHAMMAD H. ALGARNI, AHMAD ALSHAMMARI, AMJAD A. ALSUWAYLIMI, JIHANE BEN SLIMANE, MAROUAN KOUKI, AMANI KACHOUKH, REFKA GHODHBANI, TAOUFIK SAIDANI
Abstract:	This paper proposes a unified framework for defining opacity properties in both discrete and real-time systems. The framework leverages language inclusion problems to establish a common ground for expressing and comparing various opacity concepts under different observation categories. We build upon existing formalisms for opacity in Labeled Transition Systems (LTS) and Timed Transition Systems (TTS). We explain the connection between these automata models and how they are used to represent system behavior. Our framework allows for the unification of opacity definitions across these models, enabling easier comparison and analysis. Additionally, we present transformations between different opacity concepts and compile decidability results for the unified framework. Finally, we illustrate the relationships between key opacity studies through a dependency diagram.
Keywords:	Discrete Event System, Real Time System, Opacity, Verification, Decidability.
Source:	Journal of Theoretical and Applied Information Technology 31^st May 2024 -- Vol. 102. No. 10-- 2024
Full Text

Title:	DISCOVERY OF LIVER MALIGNANCE USING CONVOLUTION NEURAL NETWORK VARIANT
Author:	P.RAVI KUMAR, VISWANADHAM RAVURI, TATA BALAJI, KIRAN KUMAR KAVETI, G PRASANNA KUMAR, ELANGOVAN MUNIYANDY, N.JAYA
Abstract:	Right now, locating and detecting cancer tissue is a challenging and time-consuming procedure. Liver lesions can be segregated using cancer CT imaging to aid in treatment planning and clinical response monitoring. To segment hepatic tumours and tackle the present liver cancer issue, Mobile U-Net has been developed and is a useful tool. Liver lesion segmentation in CT scans can be utilised for therapy prediction, tumour burden assessment, and clinical outcome monitoring. This approach is a mobile device-specific modification of the U-Net architectural design. The idea is explained by the deep learning system by describing the characteristics that go into inner layer analysis and prediction and by exposing a portion of the decision-making process that pretrained deep neural network.
Keywords:	Malignance, Liver, CNN, MobileUNet
Source:	Journal of Theoretical and Applied Information Technology 31^st May 2024 -- Vol. 102. No. 10-- 2024
Full Text

Title:	THE MEDIATION ROLE OF ORGANIZATIONAL MEMORY BETWEEN KNOWLEDGE SHARING SYSTEMS AND THE COMPETITIVE INTELLIGENCE
Author:	AL-HASHEM ADEL ODEH
Abstract:	Purpose- This study aims to explore the importance of knowledge sharing systems and organizational memory as a prerequisite to competitive intelligence. Besides that, to examine the impact of knowledge sharing systems and organizational memory on competitive intelligence and the mediating role of organizational memory between knowledge sharing systems and competitive intelligence in the Jordanian banks sector in Amman. Methodology/Approach/ Design- This study uses both SPSS v26 and PLS 3.2.7 to analyze data that was collected from (580) responses that were valid for testing out of (670) questionnaires developed for this purpose Findings- knowledge sharing systems have a significant impact on organizational memory and competitive intelligence. In addition, there is significant mediation effect of organizational memory in the relationship between knowledge sharing systems and competitive intelligence. Value/ Originality- The study contribution focuses on the importance of the mediation of organizational memory between knowledge sharing systems and competitive intelligence in the banking sector in Jordan. Thus, companies have to take into account the importance role of knowledge sharing systems and organizational memory to maximize their competitive intelligence efforts.
Keywords:	Knowledge Sharing Systems, Organizational Memory, Competitive intelligence. Jordan-Amman
Source:	Journal of Theoretical and Applied Information Technology 31^st May 2024 -- Vol. 102. No. 10-- 2024
Full Text

Title:	AN EFFICIENT HOUSE ENERGY MANAGEMENT SYSTEM FOR ENERGY SCHEDULING BASED ON AN OPTIMIZED ELMAN NETWORK
Author:	ASHOK REDDYBOINA, SUBRAMANIAN RAMALINGAM2,MANISH KUMAR
Abstract:	The development of new technology has created an energy distribution system with renewable sources and a storage system for energy cost reduction. However, the houses need the Home Energy Management System (HEMS) for controlling and scheduling each device's energy to optimize energy distribution. However, the past studies do not produce satisfying results for energy optimization; therefore presented a novel Aquila-based Elman management system (ABEMS) for the energy distribution scheduling. Initially, the IHEPC dataset was trained and initialized in the system. The dataset's unwanted noise and error values were removed through preprocessing function, and the time series data were analyzed. Subsequently, the energy needed for the home appliances is calculated through the fitness function of the Aquila. Further, the energy distribution is optimized to its desired level based on the calculated value. The proposed system was implemented in Python, and the efficiency metrics were validated. Additionally, a comparative analysis is done to evaluate the improvement score.
Keywords:	Energy Resources, Feature Extraction, Aquila Optimization, Household Appliances
Source:	Journal of Theoretical and Applied Information Technology 31^st May 2024 -- Vol. 102. No. 10-- 2024
Full Text

Title:	ASSESSING THE IMPACT OF PUBLIC BLOCKCHAIN RESEARCH: A BIBLIOMETRIC APPROACH
Author:	SNEHA M KURIAKOSE, MANUSANKAR C, BENYMOL JOSE, RAJIMOL A, NIJO ANTONY, TONY CHERIAN
Abstract:	Public blockchains are open, decentralised networks that use consensus mechanisms like Proof of Work (PoW) or Proof of Stake (PoS) to validate transactions, ensuring transparency and security. This bibliometric analysis, utilising Scopus as the primary data source for its extensive journal coverage, examines the evolution of the public blockchain research scenario. The study follows a PRISMA flow diagram to identify, screen meticulously, and include relevant papers for the analysis. It focuses on trends in public blockchain studies, significant contributions, cooperative networks, research outcomes, and literature gap identification. This study utilises advanced bibliometric tools Biblioshiny and CiteSpace, providing a multifaceted view of the field's development and impact. The analysis yields significant insights into various aspects: the annual scientific production in the field, identifying the most influential authors, and discerning the most relevant sources. Furthermore, it highlights the most globally cited documents, pinpoints the most productive country in this research area, and delves into the trending topics. As a result, our findings pinpoint the dynamic nature of the field and reveal significant changes in research focus, notable contributors, and the evolution of the global collaboration pattern. Not only did the study point out the diversity and depth of public blockchain studies, but it also further identified both mature and evolving research areas. It also identified research gaps and practical consequences and proposed directions for future study and application.
Keywords:	Public Blockchains, Bibliometric Analysis, Biblioshiny, Vosviewer, Citespace
Source:	Journal of Theoretical and Applied Information Technology 31^st May 2024 -- Vol. 102. No. 10-- 2024
Full Text

Title:	SRP-EFFICIENT MODIFIED GSM CLOUD BASED WEATHER MONITORING USING WIRELESS SENSOR NETWORKS
Author:	DR BETHALA RAMYA, N.SUDHAKAR REDDY, KRISHNA SAI UJWAL KAMBHUMPATI, BATTULA SOWJANYA, V.V.RAMA KRISHNA, REDNAM S S JYOTHI, ANIL KUMAR PALLIKONDA
Abstract:	Proposed system in this paper is for record of weather conditions at a identify place and monitor the same from anywhere. Weather monitoring plays major role since weather conditions are changing dynamically. The Mobile system based wireless Telecom technology behind this will enable us to interface the things in the Internet cloud. Meteorological parameters are measured by using an automated weather station using sensors without intervention of humans. The measured parameters can be stored in a built-in data logger or can be transmitted to a remote location via a communication link. The framework manages monitoring and controlling the natural conditions like temperature, humidity, light force and CO2 level with sensors and sends the data to the remote end to store and afterward plot the sensor information as graphical insights.
Keywords:	GSM,WSN, SRP.
Source:	Journal of Theoretical and Applied Information Technology 31^st May 2024 -- Vol. 102. No. 10-- 2024
Full Text

Title:	BANYAN AGILE: A NEW APPROACH FOR MONOLITH AND MICROSERVICE DEVELOPMENT
Author:	VICTOR, MARIA SERAPHINA ASTRIANI
Abstract:	This research explores the challenges of implementing the Software Development Life Cycle (SDLC) in developing software with monolith and microservice architectures. The SDLC method serves as a framework guiding the stages of software creation, completion, and maintenance. Challenges arise when applying SDLC to software development teams transitioning from monolith to microservice architectures. Factors such as synchronization among teams, a lack of expertise in implementing specific SDLC methods, and escalating project costs become major problems . This research seeks solutions by proposing a new approach called Banyan Agile. Banyan Agile integrates SDLC principles with the flexibility of agile methods, creating a framework that can address these challenges and facilitate effective software development. The findings reveal that the implementation of Banyan Agile within PT. ASD has positively impacted collaboration among cross-functional teams and overall project productivity. Despite encountering challenges during the implementation phase, the identification of these hurdles highlights areas for improvement to optimize project outcomes.
Keywords:	Software Development, Agile, Monolith, Microservice, Strategic Planning
Source:	Journal of Theoretical and Applied Information Technology 31^st May 2024 -- Vol. 102. No. 10-- 2024
Full Text

Title:	CLUSTERING ALGORITHM FOR ELECTRICAL LOAD PROFILING ANALYSIS: A SYSTEMATIC REVIEW OF MACHINE LEARNING APPROACHES FOR IMPROVED CLUSTERING ALGORITHMS
Author:	DINE TIARA KUSUMA, NORASHIKIN AHMAD, SHARIFAH SAKINAH SYED AHMAD
Abstract:	The objective of this study is to examine a range of research studies conducted between 2017 until 2023 that focus on the analysis of Electrical Load Profiles (ELPs) using clustering algorithms within a machine learning framework. The methodology used in this research is Preferred Reporting for Systematic Review and Meta-analysis (PRISMA) framework. According to this study, it was discovered that the process of formation using the clustering algorithm can be categorized into two distinct approaches. The first approach involves the utilization of a specified number of clusters, while the second approach does not necessitate the explicit determination of the number of clusters. Additionally, it has been observed that the method employed to determine the number of clusters has a significant impact on the performance and quality of clustering, as it influences the features involved. This study explores various aspects related to clustering, including techniques for measuring the distance between data points, strategies for initializing cluster centers, approaches for reducing the dimensions of initial data, and methods for identifying and addressing outliers. The findings of this study offer insights into the various technological obstacles and emerging patterns in the analysis of ELPs, as well as investigate potential prospects for the future.
Keywords:	Clustering, Machine Learning, Load Profiles, Pattern Recognition, PRISMA
Source:	Journal of Theoretical and Applied Information Technology 31^st May 2024 -- Vol. 102. No. 10-- 2024
Full Text

Title:	DETERMINANT FACTORS OF RESIDENTIAL SOLAR PV ADOPTION IN THE PHILIPPINES: A MULTIVARIATE STRUCTURAL EQUATION MODEL ANALYSIS
Author:	TITUS JAMES R. RAGRAGIO, RIANINA D. BORRES, SATRIA FADIL PERSADA, MICHAEL NAYAT YOUNG, ARDVIN KESTER S.ONG, YOGI TRI PRASETYO, RENY NADLIFATIN
Abstract:	The objective of this research endeavor is to examine the attitudes and intentions of householders in the Philippines with respect to the implementation of Residential Rooftop Solar Photovoltaic Systems (RSPVS) and the subsequent adoption of this renewable energy technology. The research is motivated by favorable circumstances in the Philippines, which encompass abundant irradiance, favorable policies regarding solar photovoltaics, and declining expenses for systems. Gaining insight into the viewpoints of prosumers (consumers and producers) is essential for expediting the implementation of RSPVS, thereby facilitating the increased utilization of solar energy. Expanding upon prior investigations that underscore householders' profound fascination with solar technology despite significant obstacles such as exorbitant installation expenses and a dearth of comprehension, the present study endeavors to illuminate the determinants that impact prosumers' intentions to utilize RSPVS. By integrating Technophilia and Perceived Risk as supplementary constructs and utilizing a conceptual framework that combines the Technology Acceptance Model (TAM), Theory of Planned Behavior (TPB), and Unified Theory of Acceptance and Use of Technology (UTAUT), this research endeavors to identify the factors that promote or hinder the adoption of RSPVS. The research utilizes an online survey that was disseminated in prominent urban areas where Residential RSPVS is accessible. The sample consists of 239 participants, which satisfies the minimum sample size requirements. In addition to demographic information (e.g., gender, age, education, income, and sources of solar PV data), the questionnaire inquires about constructs associated with the adoption of RSPVS as outlined in TAM, TPB, and UTAUT. Using Partial Least Squares Structural Equation Modeling (PLS-SEM), the collected data are analyzed to determine the hypothesized model and to examine the relationships between variables. Technophilia, Price Value, and Social Influence are significant predictors of prosumers' intent to adopt MRSPVS, according to the findings. Significantly, Technophilia, which denotes a profound interest in novel technologies, surfaced as a crucial determinant, indicating that prosumers are driven by the appeal of solar energy. Furthermore, the research determined that Perceived Risk did not serve as a substantial impediment.
Keywords:	Residential Solar PV, Technophilia, TAM, UTAUT, TPB
Source:	Journal of Theoretical and Applied Information Technology 31^st May 2024 -- Vol. 102. No. 10-- 2024
Full Text

Title:	ENHANCING PHARMACEUTICAL SECURITY: INTEGRATING IBCDS WITH ONION ENCRYPTION
Author:	PONNADA NAGA RAMYA, DR. I. RAVI PRAKASH REDDY, DR. SUPREETHI KP
Abstract:	The pharmaceutical business must rigorously strengthen its security procedures in an era marked by increasing digitization and the enduring danger of cyberattacks. According to estimates, up to 30 percent of medical products offered in underdeveloped nations are fake. This research offers a novel approach to improving pharmaceutical security through the seamless integration of two strong security technologies: Onion Encryption and Intelligent Blockchain-Based Cryptographic Data Security (IBCDS). The pharmaceutical industry operates in a cybersecurity environment that is becoming more dynamic and changing as it is tasked with protecting a variety of sensitive data, including patient information, confidential research, and priceless intellectual property. IBCDS, which is recognised for being decentralised and tamper-resistant, provides a strong basis for protecting pharmaceutical data against intrusion and alteration. Parallel to this Onion Encryption, which takes its cues from the principles of the Tor network, provides a further layer of security by obscuring data paths and guaranteeing strict secrecy. This research offers a comprehensive and integrated security approach to improve the pharmaceutical industry's overall security posture, foster stakeholder trust, and erect an impregnable fortress against the dangers of an increasingly interconnected and vulnerable digital landscape. The proposed approach gives customers the ability to independently confirm the legitimacy of medical supplies thanks to an intuitive web interface that allows for real-time product tracking via scanned QR codes. It does this through meticulous exploration of the theoretical underpinnings of these technologies and empirical validation of their efficacy via existing literature and practical case studies.
Keywords:	Pharmaceutical Security, Blockchain technology, Cryptographic Data Security, onion encryption, Data protection
Source:	Journal of Theoretical and Applied Information Technology 31^st May 2024 -- Vol. 102. No. 10-- 2024
Full Text

Title:	ENHANCING AD HOC NETWORKS THROUGH A TRUST-BASED SECURE MULTIPATH ROUTING PROTOCOL
Author:	RAM ANUJ PRASAD, ASHOK KUMAR YADAV, SANJEEV GANGWAR, GYANENDRA KUMAR PAL, SANTOSH KUMAR YADAV
Abstract:	"Mobile Ad Hoc Networks (MANETs)" play an important role in many real-time applications, thus requiring "secured and reliable data transfer". The paper describes TBSMR, a “trust-based multipath routing protocol” proposed for the improvement of MANETs'“Quality of Service (QoS)”. TBSMR tackles issues like “congestion control, packet loss reduction”, and “malicious node detection” to enhance network performance altogether. Evaluation of the protocol is performed against previous solutions, utilizing MATLAB simulations. The performance of TBSMR is compared to others in terms of throughput, delay, and “packet delivery ratio” and is found superior. The protocol's flexibility to dynamic network environments and its ability to attach to emerging technologies are quite appealing to MANETs. In this study, secure and efficient routing protocols are promoted, and efficient operation of MANETs is achieved in different applications.
Keywords:	Mobile Ad Hoc Networks, “Trust-Based Routing”, Multipath Routing, “Quality of Service (QoS)”, Network Simulation
Source:	Journal of Theoretical and Applied Information Technology 31^st May 2024 -- Vol. 102. No. 10-- 2024
Full Text

Title:	CONVOLUTIONAL TRI BIOMETRICS: A UNIFIED APPROACH FOR IRIS, PERIOCULAR, AND FACIAL AUTHENTICATION
Author:	NALLURI PRASANTH, TRISHITHA KETINENI, SAHITHI NETHI, UDAY SAI KIRAN CHILUKURI, DR.CH.V.PHANI KRISHNA, DR.N.SRINIVASU, DR.G.PRADEEPINI
Abstract:	In response to the evolving landscape of biometric authentication, this research introduces a novel paradigm by amalgamating facial recognition with iris and periocular biometric features. Leveraging Convolutional Neural Networks (CNNs), our methodology aims to enhance the accuracy and reliability of user authentication. The research begins with the meticulous curation of a diverse biometric dataset, comprising facial images, iris scans, and periocular images. Integration of these modalities is achieved through a unified CNN architecture, ensuring a comprehensive representation of the user's identity. This model is finely tuned to extract intricate features from each biometric modality, surpassing traditional uni-modal approaches. Rigorous experimentation optimizes the model's performance, evaluating its resilience against real-world challenges such as partial occlusion and pose variations. The proposed system not only achieves state-of-the-art accuracy but also exhibits robustness against adversarial attacks and spoofing attempts. The fusion of facial, iris, and periocular biometrics enhances adaptability across diverse authentication scenarios, making it suitable for secure access control and identity verification. This research significantly contributes to the advancement of multi-modal biometric authentication, emphasizing the potential of integrating facial, iris, and periocular features through CNNs. The insights derived offer valuable implications for the development of more secure and dependable identification methods, addressing the evolving challenges within the biometrics domain.
Keywords:	Biometric Authentication, Convolutional Neural Networks (CNNs), Facial Recognition, Iris Biometrics, Periocular Biometrics, Multi-modal Authentication, User Identity, Biometric Dataset, Fusion, Uni-modal Approaches, Experimentation, Pose Variations, Adversarial Attacks, Spoofing Attempts, Secure Access Control, Identity Verification, Robustness, Authentication Scenarios, State-of-the-Art Accuracy, Authentication Systems
Source:	Journal of Theoretical and Applied Information Technology 31^st May 2024 -- Vol. 102. No. 10-- 2024
Full Text

Title:	AN INTELLIGENT NEURAL ROBOTICS FPGA DESIGN FOR PATH IDENTIFICATION
Author:	VISHAL ANDRA , S. ROOBAN
Abstract:	Autonomous robots ability to navigate various dynamic situations effectively depends on their ability to plan their paths and avoid obstacles. There are many approaches to path planning and Obstacle avoidance, and it attained an aimless way. A path without any preplanned route is called aimless. To reach its destination, the robot keeps moving randomly in one direction. Paths without goals are risky and inefficient. In addition to possibly travelling farther than is necessary, the robot might run into roadblocks. Therefore, a novel Decision Zfnet Path Planning System (DZPPS) was developed in this work. Initially, the robot's behaviour was trained using the sensed data. Then, the path was identified for the robot. After that, the robot was supposed to avoid obstacles by moving away from them when they were encountered with the assistance of the planning system. The decision function is used to avoid obstacles while driving the robot. In conclusion, measures such as path length, path efficiency, planning time, and distance to obstacles were used to determine the effectiveness of the suggested planning system.
Keywords:	Field Programmable Gate Array, Path Planning, Obstacle Avoidance, Flipflop, Lookup Table
Source:	Journal of Theoretical and Applied Information Technology 31^st May 2024 -- Vol. 102. No. 10-- 2024
Full Text

Title:	PLANT DISEASE DETECTION USING DEEP MACHINE LEARNING ALGORITHM
Author:	D.SWETHA, D. RATNAGIRI, K.VIDYA SAGAR, G.NAGA RAJU, LAKSHMI RAMANI BURRA, P. UDAYARAJU, A. GEETHA DEVI
Abstract:	The world population is increasing rapidly. In order to cater the daily needs of an individual, grains and vegetable production are imperative. This paper is focused to establish a technology support to formers and to minimize the deceases in plant. Tomato and pepper bell leaves are considered to detect the deceases. Contrast limited adaptive histogram equalization (CLAHE) is applied to improve the contrast of the leaf image before processing with machine learning algorithm. The contrast limiting is considered with clip limit 40. Bi- cubic interpolation is applied to minimize the false edge of the leaf with neighbouring tails of the leaf. The qualitative parameters like absolute mean brightness error (AMBE), mean square error (MSE), peak signal to noise ratio, mean average error (MAE) and maximum deviation (MD) are analysed. MSE values achieved less than ‘1’ indicates contrast adjustment is good. CNN Classification is applied. The decease detection accuracy with CNN is increased to 95.6 percent with increasing epochs. The accuracy Vs epoch and Loss Vs Epoch analysis is done. Optimum Tunning of hyperparameters (β1), and (β2) is done in this study. The results achieved with this approach are best fit for plant decease finding to improve the yielding rate the crop.
Keywords:	Plant leaf, CLAHE, CNN
Source:	Journal of Theoretical and Applied Information Technology 31^st May 2024 -- Vol. 102. No. 10-- 2024
Full Text

Title:	HEREDITARY APPROACH FOR FINEST DISSECTION SYSTEM FOR WEB DATA
Author:	M V GANESWARA RAO, LAKSHMI MANASA B, BALAJI TATA, DR. KARUNA ARAVA, ANE ASHOK BABU, PRAVEEN TUMULURU, N.JAYA
Abstract:	The rise of electronic data has spawned an ocean of untapped information, laying the groundwork for web mining. Simultaneously, the surge in computer technology has flooded databases with vast amounts of data, propelling the realms of Web Science and Big Data Analytics into the forefront. Web Science, an engineering process, delves into large datasets, seeking patterns amidst the chaos. Yet, extracting intrinsic structures from this vast expanse poses a formidable challenge, hindering efforts to organize them into coherent groups. Existing clustering algorithms often fall short of meeting the diverse needs of web applications. This spurred our team to pioneer an innovative algorithm, poised to offer greater applicability and resilience in this dynamic landscape. The driving force behind this research endeavor is the creation of a Machine Learning framework aimed at extracting technological insights from web data sources. The authors advocate for an Optimal Segmentation System employing a Machine Learning approach with dual objectives: firstly, to preprocess unstructured and semi-structured web documents and establish an efficient data representation structure to facilitate the application of both supervised and unsupervised techniques. Subsequently, the system prioritizes segmenting the preprocessed web data by hybridizing Genetic Approach with clustering techniques, mirroring biological evaluation processes with self-learning capabilities. Extensive experimentation has been conducted to validate the performance results of the proposed framework across various orders of magnitude, confirming its efficacy as claimed.
Keywords:	Hereditary, Machine Learning, Finest, Dissection
Source:	Journal of Theoretical and Applied Information Technology 31^st May 2024 -- Vol. 102. No. 10-- 2024
Full Text

Title:	A PROTOTYPE FOR THE DETECTION AND CLASSIFICATION OF SEISMIC EVENTS USING STA/LTA AND MACHINE LEARNING
Author:	FRIDY MANDITA, AHMAD ASHARI, M. EDI WIBOWO, WIWIT SURYANTO
Abstract:	In this study, the detection and classification of seismic events is a significant concern of this research. A volcano eruption is one of the natural disasters on Earth. Monitoring volcano activities is essential to analyzing and monitoring volcanoes before their eruption. This activity is beneficial in interpreting signals from a volcano before an eruption from the volcano can cause damage. Based on that, a tool has been developed to detect and classify volcanic seismic events. The combination of algorithm time series, which is STA/LTA and machine learning (LSTM), is being used to analyze data of seismic events. A dataset was collected from one of Indonesia's mountains during 2019 – 2021. The dataset will be classified into different classes based on the type of seismic events. Noise detection is implemented to classify true or false seismic events before continuing to detect and classify them. STA/LTA is used to remove noise signals from data seismic events. The next step is to use machine learning to classify labelling signals based on the type of seismic events. The experiments use a learning rate of 0.001 and 0.01. They show that tools can detect and classify signals of seismic events with an accuracy of around 0,70 – 0,80.
Keywords:	Seismic events, Detection, Classification, STA/LTA, Machine learning
Source:	Journal of Theoretical and Applied Information Technology 31^st May 2024 -- Vol. 102. No. 10-- 2024
Full Text

Title:	A REVIEW ON COGNITIVE BASED RANSOMWARE DETECTION USING MACHINE LEARNING AND DEEP LEARNING TECHNIQUES
Author:	MR.M.S.BALAMURUGAN, DR.V.RAJENDRAN, DR.S.SUMA CHRISTAL MARY
Abstract:	Ransomware attacks the most significant highly alluring dangers to cyber security throughout the modern era. Security software is frequently ineffective regarding extortion and zero-day spyware assaults; significant net breaches may cause considerable information loss. A competition for resources is being created by these assaults, which have become increasingly flexible and more capable of changing the way they appear. In this review article, the primary goal focusing on recent trends of ransomware detection routinely and possible directions for further research as well. This paper provides background information on ransomware, a timeline of attacks, and an overview of the virus. Additionally, it offers thorough analysis of current strategies for recognizing, averting, reducing, as well as recuperating from ransomware assaults. An additional benefit comprises an examination of studies conducted among 2016 to 2023. This gives booklover a current understanding of the most recent breakthroughs in ransomware detection and showcases improvements in techniques for preventing ransomware attacks.
Keywords:	Ransomware detection, Machine learning, Soft computing, Software Defined Network(SDN)
Source:	Journal of Theoretical and Applied Information Technology 31^st May 2024 -- Vol. 102. No. 10-- 2024
Full Text

Title:	LINEAR SUBSPACE LEARNING-BASED CLUSTER TENDENCY ASSESSMENT VISUAL MODELS FOR HIGH-DIMENSIONAL BIG DATASETS
Author:	ASWANI KUMAR UNNAM, DR BANDLA SRINIVASA RAO
Abstract:	Many new applications, including traffic image trajectories and video surveillance, require big data clustering. These applications create enormous volumes of high-dimensional data by utilizing sensors or the Internet of Things (IoT). Traditional big-data clustering techniques, such as single-pass k-implies (spkm), scaled down clump k-implies (mbkm) are broadly used to make an information segment over the enormous information. To decide the nature of the bunches covered by the huge information, they should, notwithstanding, have advance information on the group assessment. The convenience of bunching inclination for gigantic information is made conceivable by the as of late evolved examining based multi-perspectives based cosine measure visual evaluation of (group) propensity (S-MVCM-Tank). For high-layered huge information, hybrid big data clustering visual models are proposed in this study. These models address the curse of dimensionality and determine the quality of data clusters by utilizing S-MVCM-VAT and linear subspace learning (LSL). The purpose of the experimental investigation is to demonstrate how well the suggested LSL-based S-MVCM-VAT approaches perform in comparison to alternative large data clustering strategies.
Keywords:	Big Data Clustering, Clustering Tendency, Linear Subspace Learning, Multiview Points, Dimensionality Problem
Source:	Journal of Theoretical and Applied Information Technology 31^st May 2024 -- Vol. 102. No. 10-- 2024
Full Text

Title:	INTELLIGENT VEHICLE DETECTION AND CLASSIFICATION IN AERIAL IMAGERY: LEVERAGING ARFOA-ENHANCED LINEAR DISCRIMINANT ANALYSIS WITH ADVANCED VEHICLE PROPOSAL NETWORK
Author:	KRANTHI KUMAR LELLA, HANUMANTHA RAO JALLA, SANDHYA K, TEJESH REDDY SINGASANI, MOUNIKA B, SHUBHANGI JOTEPPA, SRINIVASA RAO VEMULA, RAMESH VATAMBETI
Abstract:	In the realm of computer vision, the detection of vehicles in aerial photography holds significant importance for various applications. Traditional methods rely on computationally intensive techniques with limited effectiveness in handling small objects like vehicles in large-scale aerial images. Recent advancements in deep learning, particularly R-CNNs, have shown promise but are hindered by challenges such as small object detection and the high cost of human annotation for training data. In response, this research proposes a novel system for efficient and accurate vehicle detection. Our approach utilizes a combination of deep learning techniques, including an encoder-decoder architecture for image segmentation and a hyper feature map for precise vehicle proposal generation. Additionally, we introduce the VCLDA model for vehicle classification, fine-tuned using the ARFOA algorithm. Experimental results demonstrate significant performance improvements, achieving detection rates of 84% on the Vehicle Aerial Imagery dataset, 73% on the Vehicle Finding in Aerial Imagery (VEDAI) dataset, and 64% on the German Aerospace Centre (DLR) DLR3K datasets. The proposed system has diverse potential applications, including traffic monitoring, congestion detection, intersection analysis, vehicle categorization, and pedestrian safety measures.
Keywords:	Accurate-Vehicle-Proposal-Network; Artificial Root Foraging Optimizer Algorithm; Region-Based Convolutional Neural Networks; Vehicle Detection; Vehicle Classification Based Linear Discriminant Analysis.
Source:	Journal of Theoretical and Applied Information Technology 31^st May 2024 -- Vol. 102. No. 10-- 2024
Full Text

Title:	EXPLORING ARTIFICIAL INTELLIGENCE IMPACT ON SUPPLY CHAIN’S FIRMS: A CASE STUDY OF INDUSTRIAL FIRMS IN MOROCCO
Author:	MERIEM RIAD, MOHAMED NAIMI, CHAFIK OKAR
Abstract:	Digital transformation had a significant impact on the supply chain management of enterprises. Additionally, the growing of data generated from several sensors provides opportunities to improve decision-making and profitability for firms. In this new environment, a suitable use of Artificial Intelligence technologies within the Supply Chain could lead to significant gain for firms. To ensure successful implementation for these new techniques, many factors and considerations must be taken into account. In a rapidly evolving business landscape, where efficiency and agility are paramount, organizations are increasingly turning to AI for streamline operations and enhance decision-making. The present paper aims to evaluate the implementation of Artificial Intelligence technologies, especially Machine learning techniques in industrial enterprises in Morocco, and their impact on the Supply Chain. This study delves into the strategic implementation of Artificial Intelligence (AI) in optimizing supply chain management processes. Throw a case study analysis, the research employed an in-depth interviews with industry experts, managers, and technicians to capture their experiences, perspectives, and challenges related to implementing Machine Learning solutions. Than the work offers a roadmap for organizations aiming to explore Machine Learning techniques in Supply chain under firms. However, the study also highlights challenges encountered during the implementation phase, including data quality issues, organizational resistance to change, and the need to up-skill the existing workforce to leverage AI technologies effectively.
Keywords:	Machine Learning, Artificial Intelligence, Supply Chain, Supply Chain Management, Moroccan supply Chain’s sector.
Source:	Journal of Theoretical and Applied Information Technology 31^st May 2024 -- Vol. 102. No. 10-- 2024
Full Text

Title:	A NOVEL ENSEMBLE META-FEATURES INTEGRATION TECHNIQUE FOR AUTISM SPECTRUM DISORDER DETECTION
Author:	I.SRILALITA SARWANI, D.LALITHA BHASKARI
Abstract:	Autism Spectrum Disorder (ASD) manifests as a multifaceted neurodevelopmental condition marked by difficulties in social interaction, communication, and repetitive behaviors. This paper proposes novel techniques for the detection of ASD using a combination of conventional ML algorithms and advanced ensemble techniques. Leveraging three datasets sourced from the UCI Repository, representing distinct age groups—adults, adolescents, and children, innovative approaches are introduced to enhance ASD diagnosis. After the collection of data, data preprocessing is performed. Later, the top features in each dataset are analyzed, providing insights into the most discriminative features for ASD detection. Initially, conventional ML algorithms, including logistic regression, KNN, SVM, decision trees, random forests, AdaBoost, and gradient boosting, are applied to establish a baseline for comparison. Subsequently, the effectiveness of ensemble techniques, including Bagging Meta-learner (BMA), Stacked Generalization, Stacking Classifier, and Voting Classifier, in improving detection performance is explored. Experimental findings demonstrate that the proposed ensemble techniques consistently outperform individual models across all datasets. Later, a novel ensemble meta-features integration technique was introduced, combining predictions from individual ensemble models to enhance ASD detection performance achieving higher accuracy, precision, recall, and F1-score. Finally, extended analysis was conducted to classify ASD cases into age-specific categories using ML models, achieving good results. Moreover, the techniques proposed in this research offer scalability and adaptability, suitable for implementation in diverse clinical settings. This research contributes to advancing ML-based approaches for ASD diagnosis, offering novel techniques that can potentially enhance clinical decision-making.
Keywords:	Autsim spectrum Disorder, Data Preprocessing, Ensemble Learning, Machine Learning
Source:	Journal of Theoretical and Applied Information Technology 31^st May 2024 -- Vol. 102. No. 10-- 2024
Full Text

Title:	IMPLEMENTATION OF A VOICE-GUIDED WEARABLE DEVICE FOR LOCATING THE PERSONAL BELONGINGS OF VISUALLY IMPAIRED PEOPLE
Author:	SWATI SHILASKAR, SHRIPAD BHATLAWANDE, SHRRUTI SURANJE
Abstract:	The Voice-Guided Indoor Assistance System is an innovative solution to help visually impaired individuals navigate indoor spaces. It uses Bluetooth communication and image processing techniques to locate objects and provide precise navigation instructions. The system also has a user-centric design and a unique capability to locate the device through an assistive cane. With an impressive 98% accuracy, the system ensures real-time responsiveness in diverse lighting conditions. By empowering individuals with visual impairments to navigate indoors confidently and independently, this technology marks a significant leap forward. It enhances autonomy and engagement in real-world scenarios, and the seamless integration of this system is a pivotal stride towards creating an inclusive environment for everyone.
Keywords:	Visually Impaired, Indoor Navigation, Object Localization, Computer Vision
Source:	Journal of Theoretical and Applied Information Technology 31^st May 2024 -- Vol. 102. No. 10-- 2024
Full Text

Title:	ENHANCING DISEASE OUTBREAK DETECTION: NAMED ENTITY RECOGNITION WITH FINE-TUNED DISTILBERT
Author:	MANJU JOY, DR. M KRISHNAVENI
Abstract:	Within Indias public health arena, Kerala emerges as a pioneer, often at the forefront of detecting and grappling with emerging infectious diseases. With a track record of vigilance and rapid response, Kerala has consistently been the first state in India to report the onset of various infectious outbreaks that have captured global attention. In 2018, Kerala witnessed the emergence of the first Nipah outbreak, followed by the first case of the COVID-19 pandemic in the country in 2020. Kerala's proactive surveillance mechanisms could promptly detect and notify authorities about these disasters. The first case of Monkeypox in India was also reported in Kerala in 2023, and Kerala's unwavering commitment to early detection remains unparalleled. In this scenario, a pioneering research endeavour is underway to harness the power of modern technology and computational algorithms in epidemic surveillance. Even though machines are excellent at extracting information from structured data, understanding human language and mining valuable information from unstructured text is challenging. This research presents a novel approach for detecting outbreak-related entities from amorphous text data utilizing a fine-tuned DistilBERT model with an accuracy of 96%. The model is optimized using the Optuna framework to ascertain optimal hyperparameters and ensure enhanced performance. This study advances NER methodologies in epidemiological surveillance, which is crucial for extracting relevant information from free-form text. We aim to automate the identification of disease names, affected locations, pathogen names, and population sizes from unstructured healthcare texts. Our experimental findings demonstrate that transfer learning techniques surpass baseline methods in NER tasks, mainly when training data are scarce. The proposed model is suitable for deployment in memory-constraint environments due to its capacity to operate with a reduced memory footprint and fewer resources.
Keywords:	Named Entity Recognition, LSTM, Transfer Learning, Fine tuning, DistilBERT
Source:	Journal of Theoretical and Applied Information Technology 31^st May 2024 -- Vol. 102. No. 10-- 2024
Full Text

Title:	NUMERAL SALVER UNCOVERING AND RECOGNITION USING NEURAL NETWORK
Author:	GARIMELLA BHARATHI, DR KOTTE SANDEEP, BALAJI TATA, ANURADHA CHOKKA, DR. KARUNA ARAVA, DEEVI RADHA RANI, N.JAYA
Abstract:	Salver by Numbers Acknowledgment is critical to helping the government manage automobiles effectively as the global population of private vehicles rises dramatically. However, because they are unable to recognise the number plate, slightly altered number plate formats or distinct types of number plates might present problems for currently operating NPR systems. Furthermore, the NPR system reacts quickly to changes in its surroundings. In order to appropriately tackle these problems, this study presents a novel deep learning-based NPR system. a strong NPR system that combines the three pre-processing algorithms of defogging, low-light improvement, and super-resolution. And one of the study accomplishments of this work is that the number plate is successfully recognized in a variety of situations by employing these methods. Then, by applying contours using boundary following and filtering the contours according to geographical localization and character dimensions, the number plate is effectively segmented. Character recognition is then accomplished using the EfficientNet method following de-skewing and region of interest filtering. The suggested deep learning model makes use of the ImageAI library to improve training. We use pictures of Indian license plates to assess the model's effectiveness. Character recognition achieves an accuracy of 98.78%, while number plate detection achieves an accuracy of 99.2%. In comparison to earlier approaches, the suggested strategy achieves the extensive performance. The Python platform is used for the implementation.
Keywords:	Numeral Salver, Recognition, Uncovering, Neural Network.
Source:	Journal of Theoretical and Applied Information Technology 31^st May 2024 -- Vol. 102. No. 10-- 2024
Full Text