		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
July 2022 | Vol. 100 No.14

Title:	FUNDRAISING DECISION SUPPORT SYSTEM ON INDONESIAS OIL PALM PUBLIC SERVICE AGENCY USING KIMBALL-ROSS FOUR-STEP DIMENSIONAL PROCESS AND METABASE DASHBOARD
Author:	FIRMAN FATHONI, RIYANTO JAYADI
Abstract:	Oil Palm Public Service Agency as an Indonesian local authority for supporting Oil Palm Plantation Fund Management, and directly responsible to the Ministry of Finance of Indonesia by the policies established by the steering committee concerning government programs. The organization, in this case, is required to make financial reports that aim to provide helpful information for decision making and demonstrate the accountability of the reporting entity for the resources entrusted to it. For five years, the organization relies on its data manually gathered from multiple sources of information systems and external parties, which consume a considerable amount of time and are prone to human errors. This study explains how an organization can use a Business Intelligence Dashboard to provide a quick and robust decision support system by doing automated data gathering and visualization for fast and better decision accuracy at Oil Palm Public Service Agency. This article shows that designing visually informative dashboards can help related parties understand the current situation and history.
Keywords:	Public Service Agency, Business Intelligence, Dashboard, Oil Palm, Information System
Source:	Journal of Theoretical and Applied Information Technology 31^st July 2022 -- Vol. 100. No. 14 -- 2022
Full Text

Title:	AFND: ARABIC FAKE NEWS DETECTION WITH AN ENSEMBLE DEEP CNN-LSTM MODEL
Author:	SHAYMAA E. SOROUR, HANAN E. ABDELKADER
Abstract:	The rapid expansion of misinformation in daily life has disrupted different news sources, such as social media, online news, radio and television stations, and newspapers, making it difficult to choose reliable news outlets. The potential to spread fake news (FN) to many organizations and platforms jeopardizes news credibility and causes users to abandon them. However, detecting FN entails predicting the probability that a particular news article is deceptive or not. However, most contemporary methods do not consider Arabic news and how Arabic FN (AFN) has been detected in the past decade. Therefore, research on AFN detection is beginning to receive more attention. This paper presents an Arabic FN detection (AFND) system based on hybrid deep learning (DL) model. This model includes both conventional neural network and long short-term memory (CNN-LSTM) modalities. The input dataset was prepared via discretization and normalization. Then, word vectors were included with the corrected words at a given word length as pretrained vectors on Arabic news. Due to outstanding performance, the JSO optimization algorithm was combined with the framework to automatically define the best structure for the proposed CNN-LSTM. A comparison was made between the proposed CNN-LSTM and other recent models to prove the performance of the proposed CNN-LSTM. The results indicate that the proposed CNN-LSTM offers the best performance, with an accuracy of 81.6%. The experimental results provided comprehensive improvements in the subject matter of AFND and demonstrated the potential of the proposed methodology.
Keywords:	Arabic Fake News Detection (AFND), Deep Learning (DL), Conventional Neural Network (CNN), Ensemble Learning, Optimization, Long Short-Term Memory (LSTM).
Source:	Journal of Theoretical and Applied Information Technology 31^st July 2022 -- Vol. 100. No. 14 -- 2022
Full Text

Title:	A VILLAGE MONOGRAPH INFORMATION SYSTEM MODELING: CASE STUDY MARTAPURA SUB-DISTRICT, SOUTH SUMATERA, INDONESIA
Author:	LUIS MARNISAH, HERRI SETIAWAN, JOHN RONI COYANDA
Abstract:	Village monographs are data collection carried out by the village government that is systematic, complete, accurate, and integrated in the administration of government. In Indonesia, one of the local government agencies that are obliged to update their monograph data are villages in the Martapura District, Ogan Komering Ulu Timur Regency, South Sumatra Province. The village monograph aims to make it easier for the government, community or interested parties to obtain data and information from an area, especially village data contained in the Martapura sub-district. In this case, a data and information management model is needed to manage village data which encourages researchers to build a monographic information system. This research focuses on the development of a village monograph information system model which is implemented using a software prototype. The strategy and contribution in this research is the creation of a village monograph information model that can inform all existing village potentials using the Research and Development (R&D) method, while the application prototype development uses the Unified Process (UP). The modeling system used in this study is shown in the form of a UML diagram consisting of activity diagrams, entity relationship diagrams, and use case diagrams, the results of this study are presented in the form of data in the form of village monographs in the Martapura sub-district, Ogan Komering Ulu Timur Regency which are listed in government website, so that the data presented can be reprocessed and can be updated at any time.
Keywords:	Monograph, Modeling, Information System, Unified Process, Information Model
Source:	Journal of Theoretical and Applied Information Technology 31^st July 2022 -- Vol. 100. No. 14 -- 2022
Full Text

Title:	VIDEOBEHAVIOR POSSIBLE IDENTIFICATION AND RECOGNITION OF ABNORMALITIES AND NORMAL BEHAVIOR PROFILING FOR ANOMALY DETECTION USING CNN MODEL
Author:	VENU MAJJI, DR. VANITHA KAKOLLU, MPJ SANTOSH KUMAR, K NAGA SOUJANYA, DR. B. KANTHAMMA, DR. G. BABU RAO
Abstract:	The aim of this Paper is to unravel the matter of modeling video behavior recorded in surveillance videos to be used in online normal behavior recognition and anomaly detection applications. With non-manual marking of the training data collection, a replacement architecture is made for automated behavior profiling and online anomaly sampling/detection. The subsequent are the core components of the framework supported discrete scene event detection, a compact and efficient behavior representation method is developed. Modeling each pattern employing a Dynamic Bayesian Network is employed to gauge the similarities between behavior patterns (DBN). A completely unique spectral clustering algorithm supported based on unsupervised model selection and have selection on the eigen vectors of a normalized affinity matrix is employed to get then actual grouping of behavior patterns. To detect abnormal behavior, a runtime accumulative anomaly measure is implemented, while normal behavior patterns are recognized when adequate visual evidence is out there supported a web survey. This enables the fastest possible identification and recognition of abnormalities and normal behavior. Experiments with noisy and broken data sets gathered from both indoor and outdoor monitoring scenarios show the efficacy and robustness of our approach. It’s is demonstrated that in detecting anomaly from an unseen video, a behavior model trained with an unlabeled data set out performs those trained with an equivalent but labeled dataset.
Keywords:	Dynamic Bayesian Network, Anomaly, CNN, Adaptive Video Conversion
Source:	Journal of Theoretical and Applied Information Technology 31^st July 2022 -- Vol. 100. No. 14 -- 2022
Full Text

Title:	IMPROVED SECURITY MECHANISM FOR SPECTRUM HANDOFF IN COOPERATIVE WIRELESS NETWORKS
Author:	ANAND RANJAN, O.P. SINGH, HIMANSHU KATIYAR
Abstract:	Cooperative wireless communication (CWC)is emerging as a cutting-edge technology with the goal of making opportunistic and dynamic use of unused spectrum bands. Fixed spectrum allotment by highly secure entities are results in challenging utilization of resources. In this paper, we present a unique cognitive user emulation attack (CUEA) in cooperative communication networks (CCN) that can employ during spectrum handoff to detect intruders. We next provide a safe handoff mechanism which can effectively counter quite an assault by providing a coordinating cognitive user that evaluates individual cognitive user's levels of trust predicated on its behavioral attributes. The coordinating cognitive user can effectively recognize malicious individual’s users by perusing trust factors. MATLAB simulations are used to verify the suggested mechanism's performance. The simulation results demonstrate the suggested mechanism's utility in terms of wrong authentication recognition likelihood, detection level, throughput level, and transmission time.
Keywords:	CCN, CWC, CUEA, Spectrum Handoff.
Source:	Journal of Theoretical and Applied Information Technology 31^st July 2022 -- Vol. 100. No. 14 -- 2022
Full Text

Title:	GIS-BASED FUZZY MULTI-CRITERIA DECISION MAKING METHODS: A SYSTEMATIC REVIEW AND META-ANALYSIS
Author:	PAUL AAZAGREYIR, PETER APPIAHENE, OBED APPIAH, SAMUEL BOATENG
Abstract:	Every activity must begin or conclude with a choice since decision making has become an essential element of daily life in this world. As a result, the number of contemporary decisions that require numerous considerations to be considered has skyrocketed. On the other hand, because humans are involved in the decision-making process, the multi-criteria decision-making process is burdened with incompleteness, subjectivity, ambiguity, and other fuzziness qualities. As a solution, numerous studies combined Fuzzy Set Theory with Multi-Criteria Decision Making Methods to provide effective and efficient judgments while reducing the aforementioned challenges. Regardless, geographical data-required decision making has been one of the key issues confronting the decision-making arena since the advent of multi-criteria decision making, demanding the usage of Geographic Information Systems. The current study attempted to conduct a systematic and critical assessment of around forty-nine (49) prior studies reported in academic publications on GIS-based Fuzzy Multi-Criteria Decision Making Methods throughout an eleven-year period (2011–2021). The following themes were specifically addressed: I the issue domains addressed, (ii) the research sites based on continents, (iii) the GIS-based Fuzzy Multi-Criteria Decision Making Methods employed and most generally used, and (iv) Sensitivity analysis. The findings indicated that the bulk of the research (30.61 percent) addressed the location analysis issue domain out of the 49 primary papers collected for the evaluation, while the risk assessment problem domain included the fewest studies (16.32 percent). Asia had the largest number of studies (46.93 percent), while Africa had the lowest number of publications (14.28 percent). Among the 18 primary GIS-Based Fuzzy Multi-Criteria Decision Making Methods employed, FAHP + GIS was found to be the most commonly used Fuzzy Multi-Criteria Decision Making Methods and Geographic Information Systems approaches. The study also indicated that just 17 studies (35% of the total) completed sensitivity analysis, whereas 32 studies (65% of the total) did not. Finally, we summarize the challenges and future research prospects for GIS-based Fuzzy Multi-Criteria Decision Making Methods.
Keywords:	Fuzzy, MCDM, Geographic Information Systems, FAHP, Sensitivity Analysis.
Source:	Journal of Theoretical and Applied Information Technology 31^st July 2022 -- Vol. 100. No. 14 -- 2022
Full Text

Title:	ADJACENCY MATRIX, GRAPH THEORY AND EQUIVALENCE PARTITIONING FOR MODELLING CONFORMITY TESTING OF OBJECT ORIENTED PROGRAMS
Author:	KHALID BENLHACHMI, KHADIJA LOUZAOUI
Abstract:	We present in this work an approach for testing conformity behaviours of object oriented (OO) classes. Our approach can be used to test overridden and overriding methods during the inheritance process. The key idea of our work is to use a mathematical representation for developing some algorithms of test data generation to deduce all states of conformity in the general case where behaviours of methods are not necessarily similar. Our mathematical model describes conformity contract of overridden and overriding methods during the inheritance mechanism by a graph of conformity states, adjacency matrix and equivalence partitioning. The technique of partitioning can define test cases that uncover classes of errors, thereby reducing the total number of test cases that must be developed. The second model is based on adjacency matrix and graphs of states to represent software behaviour and to simplify the test data generation. We show in this paper that the test data generation can be represented by a graph of states and adjacency matrix. Thus, it is sufficient to consider the sink vertex of graphs to check the conformity behaviour of the program under test.
Keywords:	Software Verification, Formal Specification, Conformity Testing, Robustness Testing, Valid Data, Invalid Data, Test Data Generation, Equivalence Partitioning, Inheritance, Constraint Resolution.
Source:	Journal of Theoretical and Applied Information Technology 31^st July 2022 -- Vol. 100. No. 14 -- 2022
Full Text

Title:	SPATIAL CLOAKING FOR LOCATION PRIVACY PROTECTION OF SMART HEALTH CARE SYSTEMS IN FOG COMPUTING
Author:	MAHMOUD ASASSFEH, WESAM ALMOBAIDEEN, NADIM OBEID
Abstract:	Geographic information like location data is essential for a smart health care system. Patient information data is periodically sent to hospitals or medical centers to improve healthcare services presented to patients. The location data with a timestamp can disclose the user's private data like user financial activity, medical status, lifestyle, and places frequently visited by the user. Privacy protection approaches include spatial cloaking that is used to conceal the location of the user, into a cloaking area that satisfies the user privacy requirement when using the location to get healthcare services, or when using location-based services (LBS) to get any other services. Spatial cloaking is used in many location privacy solutions, however, most of them have some disadvantages that are related to communication and computation costs. In this paper an effective spatial cloaking algorithm to preserve location privacy (LOCACY) is presented. A secure version of the A* heuristic search algorithm (SecA) has been developed to perform two essential functions: the first is to better support the proposed spatial cloaking algorithm, and the second is to enable a mobile patient to avoid infected areas while traveling between various locations. The proposed spatial cloaking algorithm outperforms rival algorithms such as bottom-up, top-down, and Aman algorithms in terms of communication and computation costs and achieves average enhancement of 56% better than the recently proposed Aman algorithm. Evaluating the secure A algorithm shows that it provides a safe path and improves the provision of privacy.
Keywords:	Smart Healthcare Systems, Fog Computing, Location Privacy, Spatial Cloaking.
Source:	Journal of Theoretical and Applied Information Technology 31^st July 2022 -- Vol. 100. No. 14 -- 2022
Full Text

Title:	INFORMATION TECHNOLOGIES AS IMPROVING TOOLS OF INCLUSIVE SOCIAL WORK PRACTICES
Author:	TETIANA LESINA, NADIIA KUZMENKO, OLENA MALOVICHKO, OLGA SHIROBOKOVA, LIUDMYLA LEVYTSKA, OLEKSANDRA SOROKINA
Abstract:	The organization of inclusive practice presupposes a creative approach and a certain flexibility of the educational system, which adapts to the individual educational needs of the individual. A unique role in preventing social exclusion is assigned to the institution of social work. Social integration is aimed primarily at the social adaptation of people with special needs to the general system of social relations within the environment into which they are integrated. The article proposes information technologies that contribute to improving tools for inclusive social work practices. The developed proposals for the use of IT technologies to improve inclusive social work practices have 4 areas: vision, tools to facilitate interaction, hearing and tools to facilitate perception. A feature of the proposed concept is using both software tools (PowerPoint, Office 365), modern IT technologies (Kahoot, Zoom, Teams, Web sites, Social networks), and physical objects (mouse, keyboard, screen).
Keywords:	Inclusive Practices, Information Technologies, Social Work.
Source:	Journal of Theoretical and Applied Information Technology 31^st July 2022 -- Vol. 100. No. 14 -- 2022
Full Text

Title:	AN EFFICIENT OPTIMIZED FRAMEWORK FOR ANALYZING THE PERFORMANCE OF BREAST CANCER USING MACHINE LEARNING ALGORITHMS
Author:	MAGDY ABD-ELGHANY ZEID, KHALED EL-BAHNASY,S.E.ABU-YOUSSEF
Abstract:	Breast cancer is a significant issue for women worldwide and a leading cause of death. This disease can be detected by differentiating malignant and benign tumors. As a result, physicians require a dependable diagnostic process for differentiating malignant from benign tumors. So automated detection of tumors is required. This research aims to introduce an optimized framework for identifying breast cancer types and predicting breast cancer recurrence using Seven Machine Learning algorithms: Logistic Regression (LR), eXtreme Gradient Boost (XGboost), Multi-Layer Perception (MLP) of Neural Network, Naive Bayes (NB), Random Forest (RF), K-Nearest neighbor (KNN) and Decision Tree (DT). We use Grid Search to optimize the machine learning algorithms. The performance of the framework was compared to determine which classifier performs the best on the Wisconsin datasets as follows Wisconsin Breast cancer (WBC) dataset, Wisconsin Diagnosis Breast cancer (WDBC) dataset, and Wisconsin Prognosis Breast Cancer (WPBC) data set. Our work presents a significant increase in cancer prediction accuracy, with the highest value being 98.3 % in the WBC dataset, 99.2% in the WDBC dataset, and 78.6% of accuracy in the WPBC dataset for cancer recurrence prediction. These results show significant progress in the area of breast cancer classification and recurrence prediction as compared to the existing state of art results of baseline machine learning models.
Keywords:	Breast Cancer, Machine Learning, Classification algorithms
Source:	Journal of Theoretical and Applied Information Technology 31^st July 2022 -- Vol. 100. No. 14 -- 2022
Full Text

Title:	RELATIONSHIP BETWEEN ONLINE SHOPPING SITES’ DESIGN AND USER EXPERIENCE USING A SURVEY
Author:	NUR FARAHA MOHD NAIM, SHARON ELIANNA MOHD ADZLAN, ROZITA ISMAIL, NUR AINNA RAMLI, HO CHI KIAN, ASLINA BAHARUM
Abstract:	The functionality of a products design should always be prioritized. A good design in a product will affect the experience of users and their emotions simultaneously. It is well-known that a good design will influence users with positive emotions (calmness, happiness, satisfaction), while a bad design will influence users with negative emotions (anxiousness, fear, unsatisfied). The emotions felt by users may also impact their decisions and thus, affect the shopping site’s business at the same time. It has been said by many psychological scientists that emotions influence one’s life decisions. Even in human-computer interaction, users’ experience could trigger a human emotional response caused by certain factors which lead to their decision to use. Therefore, this research is conducted to examine how negative emotions may influence users' experience in another domain, online shopping sites and assess the possibility of a correlation between the online shopping sites’ design and the human emotional response. An experimental research technique is used. A survey was conducted on 31 respondents to gain insight into the effectiveness of this research to evaluate the user’s Key Performance Index (KPI) of their shopping process and improve the design of online shopping sites; after a thorough analysis of the completed survey results in negative emotions on users as they shop on online shopping sites. Two analytical techniques, Descriptive Statistics Analysis (Frequencies) and Pearson’s r and Scatter Plots, were then used as metric measurements to analyze the user’s feelings data obtained from the respondents. Three prominent feelings, such as engagement, boredom, and frustration, were selected as user’s KPI data level for Descriptive Statistics Analysis (Frequencies). The other analytical technique is Pearson’s r and Scatter Plots to evaluate the correlation strength between the six scales' feelings response toward the design. Results show that the existing online shopping sites have somehow mildly triggered the KPI index of boredom and frustration among the respondents involved. Even so, the KPI index of engagement is still on the positive side, and according to the six scales’ feelings results, certain design components correlate with user’s feelings and some are not correlated. Thus, to levitate the positive emotion is to consider giving more attention to the design component.
Keywords:	Negative Emotions, Online Shopping, User Experience, Interface Design, Emotion-based Experience
Source:	Journal of Theoretical and Applied Information Technology 31^st July 2022 -- Vol. 100. No. 14 -- 2022
Full Text

Title:	A PREDICTION ON EDUCATIONAL TIME SERIES DATA USING STATISTICAL MACHINE LEARNING MODEL -AN EXPERIMENTAL ANALYSIS
Author:	VANITHA.S, JAYASHREE.R
Abstract:	Prediction using time-series data is a vital part of machine learning because it keeps the temporal information of historical data for forecasting. Time series analysis is extensively used in all sectors wherever the data is populated and estimated based on timing such as seconds, minutes, hours, days, months, quarterly, half-yearly, and yearly. However, the model accuracy relies on the number of observations (data), consistency, and the consequence of data. The contribution of this paper is finding the trend of an educational institution enrollment in the upcoming year using the statistical machine learning mode1. Then, a detailed study has conducted to find the capability of the statistical model observed in various scenarios to handle time-series data. The study reveals the factors (Model fitness, Best forecasting duration, Impact of Train/Test ratio in precision) affecting the model accuracy of the statistical algorithms. This work also fulfills the research gap where less work has conducted in year-wise cyclic data without any trend. The methods used for this experiment are Auto-Regressive Integrated Moving Average (ARIMA) and Simple Exponential Smoothing (SES) technique. Finally, the two models are compared and the research objectives are discussed with the experimental result. Mean Absolute Percentage Error (MAPE) and Root Mean Squared Error (RMSE) metrics are used for assessing the model precision. The experimental result proves that the SES model provides better performance than ARIMA and both models are executed with their own merits and demerits.
Keywords:	Time Series, ARIMA, Simple Exponential Smoothing, Student Enrollment Prediction, Factors Affecting the Model
Source:	Journal of Theoretical and Applied Information Technology 31^st July 2022 -- Vol. 100. No. 14 -- 2022
Full Text

Title:	COMPARATIVE ANALYSIS OF MAXIMUM POWER POINT TRACKING METHOD USING SLIDING MODE AND P&O CONTROLLERS
Author:	SANA MOUSLIM, M HAND OUBELLA, MOHAMED AJAAMOUM, EL MAHFOUD BOULAOUTAQ, MOHAMED BENYDIR, KAOUTAR DAHMANE
Abstract:	The power delivered by a solar photovoltaic generator (PVG) strongly depends on the level of irradiance G, temperature T of cells, total or partial shading but also the nature of the fueled load. The PPV-VPV power characteristic of the PVG has a maximum power point (MPP) corresponding to the optimal operating point. Since the position of the MPP depends on the level of irradiance and the temperature of the cells, it is never constant over time. Therefore, a control strategy is requisite to extract maximum power from solar panels under all operating conditions. The objective of this work is to design a MPPT controller based on sliding mode controller (SMC) that is applied to a buck-boost converter in order to achieve an optimal PV module output voltage. The proposed MPPT controller using SMC has been developed so that the operating point converges to the optimum operating point. The validation of the proposed controller is shown by MATLAB/SIMULINK simulation. The results confirm the effectiveness of the sliding mode control MPPT under the parameter variation environments. Moreover, a comparison analysis of the proposed SM controller and classical MPPT algorithm using Perturb-and-Observe method has been designed for the same PV power system in order to evaluate the robustness and stability against parameter uncertainties for the two proposed controllers.
Keywords:	Photovoltaic; Buck-Boost converter; MPPT; P&O; Sliding mode control; MATLAB/Simulink.
Source:	Journal of Theoretical and Applied Information Technology 31^st July 2022 -- Vol. 100. No. 14 -- 2022
Full Text

Title:	DEEP LEARNING BASED MODEL FOR TABLE DETECTION CONTENT AND LAYOUT ANALYSIS IN COMPRESSED DOCUMENT IMAGES - A COMPREHENSIVE APPROACH
Author:	KAVITA V. HORADI, DR. JAGADEESH PUJARI, NARASIMHA PRASAD BHAT
Abstract:	Nowadays, the digital data is generated abruptly in the form of digital documents. Generation of large volumes of digital data giving rise to the big data problems has invited various problems which research community need to address. Exponential increase in the capacity of ‘Big-data’ containing images, textual information, audios and video content has paved a way to many challenges in processing because of an unstructured content. Due to large number of indexing and analyzing these images becomes a challenging issue. As there are various compression techniques available worldwide, these document images may undergo any compression before storage or transmission due to space and bandwidth issues. Once a document image is compressed it generates a compressed document image (CDI) which will have complexity in processing due to the loss of vital information present in it. Moreover, recognizing the layout of these documents is an important stage for various applications thus document layout analysis and recognition is considered as a promising solution for various computer vision based applications. Currently, deep learning schemes are widely adopted and comparative analysis has proven the accuracy of deep learning schemes. However, the accuracy of these systems is affected due to unstructured form of data. To overcome this issue, we present a novel scheme for layout and content equivalence analysis in compressed domain. The proposed approach uses a deep learning technique for detecting a table and faster RCNN based model for identifying the ROI. Moreover, this model incorporates the contextual information to improve the detection accuracy corresponding to each label in ROI. The proposed approach is tested by using publically available PubLayNet dataset. the average precision of PubLayNet dataset is obtained as 97.50%, F1-socre for DocBank is obtained as 97.09% and 96.55 mAP for DocBank. The comparative analysis proves that the proposed novel method attains better performance when compared with existing schemes.
Keywords:	Deep Learning; Compressed Document Images (CDI); Table Detection; Layout Analysis; Content Analysis; Faster RCNN, Datasets.
Source:	Journal of Theoretical and Applied Information Technology 31^st July 2022 -- Vol. 100. No. 14 -- 2022
Full Text

Title:	CREDIT CARD FRAUDS SCORING MODEL BASED ON DEEP LEARNING ENSEMBLE
Author:	SHUMUKH AL-FAQIR, OSAMA OUDA
Abstract:	Credit card frauds can result in substantial financial losses, particularly when fraudulent transactions have large values. Thus, it is essential to detect fraudulent transactions prior to their authorization by card issuers. Most conventional fraud detection systems are based on machine learning models. Recent studies explored utilizing deep learning (DL) models to detect fraudulent transactions efficiently. However, such studies depend merely on a single DL model. In this paper, we present various deep learning and ensemble methods for detecting credit card fraudulent transactions. The main motivation behind this presented work is to contribute toward reducing both missed frauds and false alarms, where our contribution in this work lies specifically in combining the resulting scores of three different distinct DL models, namely, convolutional neural networks (CNN), autoencoders (AE), and recurrent neural networks (RNN). Experiments on a public credit-card dataset demonstrated that, for the single DL-based models, AE has the best validation accuracy (93.4%) compared to CNN (91.4%) and RNN (91.8%). For the ensemble results, the validation accuracy (94.9%) was superior to all the three implemented DL-based models.
Keywords:	Deep Learning, Convolutional Neural Networks (CNN), Auto Encoders (AE), Recurrent Neural Networks (RNN), Ensemble Learning.
Source:	Journal of Theoretical and Applied Information Technology 31^st July 2022 -- Vol. 100. No. 14 -- 2022
Full Text

Title:	BLOCKCHAIN TECHNOLOGY AND ELECTRONIC SUPPLY CHAIN MANAGEMENT (ESCM)
Author:	SAMIA FEKIR, MOHAMMED AREF ABDUL RASHEED, BILLAL CHIKHI, FATIMA ALMARHOON, RAWAN AL MASHIKHI
Abstract:	The study tried to examine to what extent technological, organizational and environmental factors influence the employees’ intention to use block chain technology in the supply chain system of their companies. The study adopted the Technological, Organizational, and Environmental framework as a theoretical base. More, the structural equation modelling technique implemented to examine the proposed hypotheses. Moreover, the findings revealed that only technological and organizational factors have significant effect on employees’ intention to use block chain technology in the supply chain system of their companies. Companies are highly advised to provide more support to the employees in general and those with technological competences specifically as they are expected to play a vital role in supporting the decision of block chain technology use in e-supply chain system of their companies. Furthermore, more investment in technological infrastructure as technology have strong effect on organization to accept and use block chain technology in e-supply chain system.
Keywords:	Blockchain Technology, Cryptocurrency, Electronic Supply Chain Management, Structural Equation Modeling
Source:	Journal of Theoretical and Applied Information Technology 31^st July 2022 -- Vol. 100. No. 14 -- 2022
Full Text

Title:	TOWARD A CONCEPTUAL METAPHORS MODEL TO SUPPORT THE IOT-DATA ANALYSIS PROCESS
Author:	RAKI YOUNESS, MARZAK ABDELAZIZ, GHRAR ADIL
Abstract:	The analysis and processing of IoT data is related to its changing behavioral nature. Thus, describing the behavior of data to infer how it is presented can reduce the load on processing and analysis resources for more efficient performance. Therefore, the adoption of a particular technology has become an insufficient measure, which necessitates the transition to the mode of research in other related fields. In this paper, we present a conceptual metaphors (CMS) architecture to support the IoT data consistency process. This model aims to create metaphorical relationships between data that connect entities to describe and produce meaning, based on the study of the possibility of combining concepts of metaphor and semantic approach. It can provide suggestions on how to improve the use of resources as a service to deal with massive IoT data. Moreover, adopting this metaphorical mechanism can enable developers and designers to design and re-architecture adapted to the nature of this data. The results of the analytical study show that our proposed model has high descriptive features that make it more efficient than other existing models.
Keywords:	IoT-data, Cloud, conceptual metaphors, Sensors networking, Resources overload, Big data.
Source:	Journal of Theoretical and Applied Information Technology 31^st July 2022 -- Vol. 100. No. 14 -- 2022
Full Text

Title:	HYBRID ENSEMBLE MODEL FOR FAKE NEWS DETECTION
Author:	M.S.H. BASRI, N.H. ABD RAHIM
Abstract:	The increasing number of fake news spread to give a negative impact to many parties including society. Hence, to counter the high volume of fake news dissemination problems, numerous machine learning techniques have been used for automated fake news detection. Thus, this study implements the machine learning algorithms by developing a hybrid ensemble model for the classification of fake news. The primary objective of this work is to increase the performance of the machine learning model by developing the hybrid ensemble model to assist in the automatic detection of fake news in classifying the labelled news. This paper focuses on implementing the model with unorganized data in textual forms of Malay language news. A few performance metrics such as recall, accuracy, precision, and f-score will be used to measure the performance of the proposed model. The model has successfully increased the accuracy of fake news detection with 75% score.
Keywords:	Automated Detection System, Fake News, Hybrid Ensemble Model, Machine Learning, Malay Language,
Source:	Journal of Theoretical and Applied Information Technology 31^st July 2022 -- Vol. 100. No. 14 -- 2022
Full Text

Title:	FEEMD AND GWO METHODOLOGY FOR FLOOD EARLY WARNING PREDICTION MODEL
Author:	NOOR HAYATI MOHD ZAIN, NORAFIDA ITHNIN
Abstract:	Flood disasters are natural hazards that cause many great losses either in terms of lives, property, and even the structure of the earth's surface. Investigations on this topic have become one of the ongoing studies because it has a great impact on the environment and community life. This study also highlights the improvement of flood warning measurement methods to ensure that the adverse effects of flood disasters can be controlled better than before. This paper will present a more efficient method to model flood early warning prediction in Malaysian districts particularly. This study focuses on the use of Fast Ensemble Empirical Mode Decomposition (FEEMD) to decompose selected rainfall dataset. Furthermore, the Gray Wolf Optimizer (GWO) was used as an optimization approach to optimize between FEEMD hybrids with Artificial Intelligence models to find the most accurate flood early warning prediction model. The study also aims to improve the process flow of the flood early warning prediction system delivered to flood victims by ensuring that they are able to prepare for the consequences of impending flood disasters in their areas of residence.
Keywords:	Flood Warning Prediction, Natural Disaster, Monsoon Flood, FEEMD, GWO
Source:	Journal of Theoretical and Applied Information Technology 31^st July 2022 -- Vol. 100. No. 14 -- 2022
Full Text

Title:	DATA MINING AND FEATURE EXTRACTION TECHNIQUES FOR OPINION MINING ANALYSIS IN TWITTER
Author:	A. AL FIRTHOUS, P.ARUL
Abstract:	By keeping and gaining customers, every small and large firm hopes to increase earnings and sales. It is vital to estimate future sales in order to improve sales. Aim of this work is to find the product is liked and recommended by the customer or not. To retain customers, product sellers must assess online customer demand in order to manage product, brand, and inventory levels. To improve customer service, online retailers or sellers must first understand the needs of their customers. It's also important to think about future sales. The proposed opinion mining on Amazon products is based on various sorts of real-time data obtained from Twitter. Lexical features, sentiment features and pragmatic features have also been used to improve the classification of review opinion features. After preprocessed the collected real time reviews have been involved in the process of feature extraction. Finally positive and negative opinions about the Amazon products have been classified by the proposed classification model. The proposed work findings and results states whether the Amazon shopping customers likes and recommend the product or not recommend. If the positive opinionated reviews are higher, then the customers like the product and recommend the product to purchase. If the negative opinionated reviews are higher, then the customers not like the product and not recommend the product to purchase. This research aims to aid clients in making purchasing decisions.
Keywords:	Opinion Mining, Rule Based, Lexicon, Pragmatic, Prediction, and Classification)
Source:	Journal of Theoretical and Applied Information Technology 31^st July 2022 -- Vol. 100. No. 14 -- 2022
Full Text

Title:	CNN-GRU-BASED HYBRID APPROACH FOR COVID-19 DETECTION THROUGH CHEST X-RAY IMAGES
Author:	D. RAGHU, HRUDAYA KUMAR TRIPATHY
Abstract:	All intelligent devices using the current world generate a massive amount of data every day. Medical field also facing challenges to classify large datasets to trace out disease cause. Automatic disease detection frame work is suitable to the current world. Programmed discovery of disease has now become an important problem in clinical science as a result of rapid population growth. A programmed location of the disease structure helps professionals identify infections and provides accurate, predictable and rapid results and reduces the rate of passage. Covid (Coronavirus) has probably had the latest and most intense illnesses and has spread all around the world. As the quickest indicative alternative, a mechanized recognition framework should therefore be used to prevent the spread of coronavirus. Coronavirus disease (COVID-19) is a contagious infection caused by a newly identified coronavirus. Automatic Disease Identification is still challenging thing at present situation. COVID-19 outbreak is the most recent risk to global health. There are currently only few COVID-19 collection is available for protecting privacy, while large data sets for CXRs are available. COVID-19 biomedical papers are growing rapidly at the same time, including reports on radiological findings. These massive data were easily identified using convolutional neural networks by classification and object identification. CNN is a popular technique for object recognition using different CNN algorithms. This paper proposes an amalgam of Deep Learning approach based on Convolutional Neural Network-CNN to extract features from Chest X-ray images, it also proposes Gated Recurrent Units GRU which can be used for the purpose of classification of Chest X-ray images. The results generated by our proposed model are as 0.94, 0.96, and 0.97 in terms of Precision. This model can aid the doctors to perform early detection of Covid-19 and makes world as Covid Free.
Keywords:	Covid-19, Detection, Chest X-ray image, Convolutional Neural Network, Gated Recurrent Units, Classification
Source:	Journal of Theoretical and Applied Information Technology 31^st July 2022 -- Vol. 100. No. 14 -- 2022
Full Text

Title:	METHODS OF SHORT TERM ELECTRICITY DEMAND FORECASTING
Author:	G.TAGANOVA, J.TUSSUPOV, A.ABDILDAYEVA, S.SERIKBAYEVA, ZH. SADIRMEKOVA, G.AZIEVA, D.MAMATAYEVA, M.TOLGANBAYEVA
Abstract:	This article discusses the short-term forecasting of electricity consumption. Methods of smoothing the daily schedule of power consumption are considered. The possibility of using one of the methods of smoothing power consumption was analyzed in Python. The proposed method is applicable for the subjects of REM in order to approximate the retrospective data of electricity consumption. The relevance of the work is due to the demand of the subjects of the wholesale electricity and capacity market (REM) for ways to build short-term forecasts of electricity consumption in order to improve the quality and accuracy of the predictive model. From the conducted research, it was revealed that the adaptive Holt-Winters smoothing method is optimal for making short-term forecasting for the day ahead.
Keywords:	Forecasting; Power Consumption; Data Analysis, Python.
Source:	Journal of Theoretical and Applied Information Technology 31^st July 2022 -- Vol. 100. No. 14 -- 2022
Full Text

Title:	BENFORDS LAW AS A TOOL IN DETECTING FINANCIAL STATEMENT FRAUD
Author:	ARYANTI WARDAYA PUPOKUSUMO, BAMBANG LEO HANDOKO, WILLY, RICKY, EDWIN HENDRA
Abstract:	Currently, fraud related to financial statement fraud is increasingly happening. The parties who are harmed are investors and creditors who make decisions based on the financial statements. Benford's Law is here as one of the tools to detect fraudulent financial statements. Benford's law is the study of the frequency of the principal digits contained in numerical data. It is also commonly used in predicting the occurrence of numbers in numerical data, including auditing financial statements. When an Auditor chooses a method of detecting fraud / material misstatement of data, he should first consider which types of accounts that may be analyzed by the Benford method are expected to be effective or not, While most of the accounting data sets related to the Benford distribution are in accordance with the Benford distribution because digital analysis is only effective when applied to the appropriate data set. Auditors need to consider in advance the expectations for the use of the Benford method distribution before conducting digital analysis. The purpose of this study is that we want to demonstrate the effectiveness of the Benford Law method in assisting the auditing process. Our research result show that Benford Law method is still effective in helping auditors detect fraud. It can be seen from the results that we get we can assess the size of the anomaly, the risk of fraud, and changes in deviation so that it gives rise to indications of fraud in Total Assets, Total Liability, and Total Equity by using the Benford Law method, namely First Digit Test, First Two-Digit Test, and Chi-Square Test.
Keywords:	Benford Law, Data Fraud, Audit, Transportation, Accounting
Source:	Journal of Theoretical and Applied Information Technology 31^st July 2022 -- Vol. 100. No. 14 -- 2022
Full Text

Title:	A HYBRID PLACESNET-LSTM MODEL FOR MOVIE TRAILER GENRE CLASSIFICATION
Author:	DAYOU JIANG
Abstract:	Video analysis technology has always been an essential branch of computer vision. Analyzing the movie genres is beneficial to pushing relevant and exciting content to target customer groups to achieve precision marketing. There are some researches on movie trailers to classify movie genres. However, most of them are based on movies' auditory and visual content using various machine learning models or neural network models for classification. This paper considers the features learned using scene-based neural network models in movie genre classification. This paper proposes a hybrid PlacesNet-LSTM (long short-term memory) model for movie trailer genre classification. To compare the performance, the paper also studies two schemes using various video and audio features based on multiple machine learning models and LSTM, respectively. The experimental results show that the PlacesNet-LSTM model on scene recognition achieves the best classification performance in various combinations.
Keywords:	Movie, Genres Classification, Machine Learning, Long Short-Term Memory Networks, Scene Recognition
Source:	Journal of Theoretical and Applied Information Technology 31^st July 2022 -- Vol. 100. No. 14 -- 2022
Full Text

Title:	COMPREHENSIVE ANALYSIS ON INTELLIGENT DEEP LEARNING BASED APPROACHES FOR HEART BEAT DETECTION
Author:	SAAD ALMUTAIRI, MANIMURUGAN S, MAJED MOHAMMED ABOROKBAH, NARMATHA C, SUBRAMANIAM GANESAN, RIYADH A ALZAHEB, HANI ALMOAMARI
Abstract:	In recent years, the research study on heartbeat detection has been increased, which is more essential in medical and sports-related applications. These analyses help to find most heart disorders by examining the electrical signal of the heartbeat that produced with distinct unique cardiac tissues located in the heart of the body. Recently, numerous works have been developed to generate class labels based on automatic heartbeat classification techniques. More importantly, Deep Learning (DL) approaches used in recent times to optimize the functionality of traditional heartbeat methodologies. With this motivation, this study analyses the DL methods of ECG-based automatic heartbeat abnormalities detection through analyzing the ECG signal pre-processing to improve the quality, heartbeat segmentation techniques to identify the target region, feature extraction methods to reduce complexity of classifier by reducing the number of resources, and different DL based classification algorithms to generate class label for identifying the heartbeat. Finally, this analysis focus on the difficulties that DL models encounter and suggest some potential future directions. The results observed from various studies clearly show that classification performance improves even when using datasets with limited sample size. This study suggests that further attention should be paid to enhancing the generalizability of DL models used to analyse ECG signals, particularly by extracting more significant sample datasets.
Keywords:	Heartbeat Measurement; Electrocardiogram Signals; Cardiovascular Diseases and Deep Learning
Source:	Journal of Theoretical and Applied Information Technology 31^st July 2022 -- Vol. 100. No. 14 -- 2022
Full Text

Title:	INTERACTIVE GEOINFORMATION MAP OF DEGRADED PASTURES OF KAZAKHSTAN WITH DIFFERENT DEGREES OF DEGRADATION AND MEASURES FOR THEIR MANAGEMENT
Author:	SAGYNBAY KALDYBAEV, KUANYSH ZHOLAMANOV, KENZHE YERZHANOVA, AIGUL BEKETOVA, ZHAINAGUL ERTAEVA, BAGDAT RUSTEMOV
Abstract:	This paper presents an assessment of pasture areas of Kazakhstan with different degrees of degradation of desert, foothill semi-desert (vertical zoning), semi-desert (latitudinal zoning), dry-steppe, steppe and forest-steppe zones using GIS technology based on field survey data with indicators of physical (soil) and biological (botanical) indicators and medium and low resolution satellite data. The aim of this work is to develop an interactive geo-information map of pasture degradation in Kazakhstan. To determine pasture degradation, the remote sensing (RS) method was used with all available mapping material and satellite data from Landsat 8, Sentinel 2, Modis TERRA, which resulted in analysis of vegetation indices NDVI, SAVI, BareSoilIndex, SalinityIndex, Top-SoilGrainSizeIndex for 33 monitoring areas in 12 oblasts in the context of degradation contours. A database for all soil-geographical zones was compiled and an interactive geo-information map of the degree of pasture degradation for the studied soil-geographical zones of Kazakhstan at a scale of 1:750 000 was developed, which includes soil, botanical characteristics and measures for their management and improvement.
Keywords:	Monitoring, GIS Technology, Remote Sensing of the Earth (RSE), Physical Indicator, Biological Indicator.
Source:	Journal of Theoretical and Applied Information Technology 31^st July 2022 -- Vol. 100. No. 14 -- 2022
Full Text

Title:	CLASSIFICATION PERFORMANCE FOR CREDIT SCORING WITH ENSEMBLE METHOD APPROACH
Author:	ABBA SUGANDA GIRSANG , CHRISTOPHER EDMOND
Abstract:	Credit scoring is an important part in controlling risk in financial companies. With the high number of non-performing loans, the assessment of potential new customers in financial companies has become a major focus of the financial industry. High accuracy credit scoring system can give better predictions on new customers and can change the company's economic growth and for better capital. This study uses a real world dataset, where data is obtained directly from a financial company and will be used to feed a random forest model to differentiate between good and bad potential new customers. This contribution of this research is to improve single model with real dataset by using ensembled bagging, bootstrap aggregating. Two methods are implemented, random forest and neural network to see the performance of ensembled using bootstrap aggerating. The output accuracy from the final model of the ensemble methods that resulted from the voting will be compared with the original unmodified single model and another model with similar architecture.The result shows that the modified multiple model surpasses the unmodified single model in terms of accuracy with a tradeoff on duration in the process.
Keywords:	Bootstrap Aggregation, Credit Scoring, Ensemble Methods, Random Forest
Source:	Journal of Theoretical and Applied Information Technology 31^st July 2022 -- Vol. 100. No. 14 -- 2022
Full Text

Title:	HARDWARE IMPLEMENTATION OF GALOIS FIELD MULTIPLICATION FOR MIXCOLUMN AND INVERSEMIXCOLUMN PROCESS IN ENCRYPTION-DECRYPTION ALGORITHMS
Author:	RAGIEL HADI PRAYITNO, SUNNY ARIEF SUDIRO, SARIFUDDIN MADENDA, SURYADI HARMANTO
Abstract:	This article described the Advanced Encryption Standard (AES) mixcolumns (MC) and inverse mixcolumns (IMC) process based on the multiplication of Galois Fields (GF 28). Multiplied the original data and Matrix MC will have resulted an Encryption. In decryption, the IMC transformation method was same as which used during encryption, where the input for decryption was the encryption matrix array data and IMC matrix. The output which was generated in the decryption was the original matrix array data. The transformation of MC and IMC was conducted using two methods; they were based on the multiplication of Galois Field (GF28) and based on tables L and E. The method in this article had been applied to MATLAB software and implemented in hardware using Field Programmable Gate Array (FPGA) device. In the implementation of hardware, it required 28 slice LUT, 28 LUT-FF and 24 bonded IOB with process time (delay) of 2.236 ns.
Keywords:	AES, FPGA, Galois Field, Inverse Mixcolumns/Mixcolumns.
Source:	Journal of Theoretical and Applied Information Technology 31^st July 2022 -- Vol. 100. No. 14 -- 2022
Full Text

Title:	USING MACHINE LEARNING TO PREDICT THE SENTIMENT OF ARABIC TWEETS RELATED TO COVID-19
Author:	ESLAM AL MAGHAYREH
Abstract:	The world has suffered a lot due to the spread of the COVID-19 virus. The world health organization (WHO) declared the COVID-19 pandemic a global emergency. The governments and healthcare officials everywhere were fighting to control the spread of this pandemic. Meanwhile, a considerable amount of social media data (e.g., tweets) related to COVID-19 is being generated continuously. In this paper, we will build a model that can identify the sentiment of Twitter data related to COVID-19 using machine learning. We will focus on analyzing Arabic language tweets to determine people’s opinions, feelings, and status on the impact of COVID-19. The main objective of this research is to support efforts to study the impact of the COVID-19 pandemic on society. To achieve this objective, we have prepared a dataset of Arabic tweets related to COVID-19 and manually classified the tweets in the dataset. Then we have used machine learning to develop an approach to assess people’s feelings about COVID-19. This approach can help the government and healthcare officials to identify any negative and positive aspects of this crisis to improve their response to similar future crises.
Keywords:	Sentiment Analysis, Machine Learning, Text Analysis, Natural Language Processing, Data Science
Source:	Journal of Theoretical and Applied Information Technology 31^st July 2022 -- Vol. 100. No. 14 -- 2022
Full Text

Title:	MATHEMATICAL AND COMPUTER MODELING OF CRITICAL AREAS OF LOSS OF STABILITY OF COMPLEX SYSTEMS
Author:	KULSHAT AKANOVA, ASSEM MYRKANOVA, ANAR ZHUMAKHANOVA
Abstract:	In recent years, the pace and scale of the development of minerals and energy resources in mountainous areas has been increasing, so various ground and underground structures are being built and operated. Increasing anthropogenic load on the biosphere and technosphere leads to the intensification of natural and man-made disasters. In addition, the occurrence of a catastrophe of the first type can provoke the manifestation of a catastrophe of the second type, and vice versa. Using ideas about catastrophic phenomena and the dynamics of their development, rock mechanics and methods of mathematical and computer modeling, it became possible to apply theoretical knowledge in practice. The issues of forecasting, preventing and minimizing events leading to loss of life and economic damage are among the most relevant today [1]. The article shows a study of the stress-strain state of an underground tunnel, which is affected by the volumetric weight of a rock mass. Deformation and displacement of mountain ranges can lead to the loss of stability of the technical structure and to its collapse, and thus cause a man-made disaster. The analytical and numerical solution of the problem will allow engineers and specialists in the field of mining to improve the safety and reliability of underground structures and take preventive measures to prevent the risk of their destruction. In this article, formulas are obtained for determining the components of stresses, deformations and displacements in an untouched rock mass, as well as on the contour of a constructed underground structure at bifurcation points. With their help, zones of high concentration of stresses, deformations and displacements and changes in the nature of their distribution are identified, which signal a critical state of the system, in which it can lose stability. As an example, an underground working of an elliptical profile is considered.
Keywords:	Stress, Deformation, Collapse, Bifurcation, Catastrophe.
Source:	Journal of Theoretical and Applied Information Technology 31^st July 2022 -- Vol. 100. No. 14 -- 2022
Full Text

Title:	BREAST CANCER DIAGNOSIS AND PROGNOSIS USING STACKING ENSEMBLE TECHNIQUE
Author:	NOURHAN M. SWELAM, AYMAN E. KHEDR, HEND AUDA
Abstract:	Breast Cancer is the most common type of cancers in Egypt, early diagnosis can help to lower the risks. For many physicians, predicting a cancerous tumor remains a challenging task also deciding which treatment plan would help the most. The availability of new medical technologies and the massive amount of patient data had motivated the basis of emergence of new strategies in the prediction and detection of cancer. Data mining analysis and Machine Learning (ML) techniques can help to develop tools that can be used as effective mechanism for early diagnosis and prognosis of breast cancer, which will greatly enhance patients' survival rate. The main objective of this paper is to compare between the performance of supervised learning classification algorithms and the performance of combination of these algorithms using stacking ensemble learning approach in terms of the classification accuracy, precision, recall and ROC. We conducted the experiments on breast cancer dataset collected from University of California, San Francisco. The results demonstrate that the proposed stacking ensemble learning model outperforms individual algorithms.
Keywords:	Breast cancer, Stacking, Classification; J48, KNN, Naïve Bayes, Support vector machine
Source:	Journal of Theoretical and Applied Information Technology 31^st July 2022 -- Vol. 100. No. 14 -- 2022
Full Text

Title:	DE-ANONYMIZATION OF THE USER OF WEB RESOURCE WITH BROWSER FINGERPRINT TECHNOLOGY
Author:	EVGENY KARPUKHIN, VADIM SHARMAEV, ANTON PROPP
Abstract:	The paper highlights the main characteristics of the user, which can be used in the formation of the browser fingerprint, revealing their features. De-anonymization of the user can be used to create individualized advertising campaigns that match the interests of the person, to improve systems for recommending content (for example, articles, videos and music), for secure authentication, collecting statistics about site visitors and analytics. The article also presents other possible scenarios for applying the technology. The methodology presents three possible scenarios: cross-browser solution, maximum amount of data and high accuracy. For each of them, the most appropriate array of user characteristics used to form the fingerprint is chosen, and examples of the JavaScript script are demonstrated. The disadvantage of the technology is the fact that when we change the value of one of the analyzed parameters, the entire output data block also changes. The solution to this problem is to choose the optimal sensitivity threshold. Calculated the optimal sensitivity threshold depending on the number of analyzed parameters, we give examples of its use to determine whether to consider the web service user as a repeat visitor or a new user.
Keywords:	User de-anonymization, Browser fingerprint, Device fingerprint, Information security, JavaScript.
Source:	Journal of Theoretical and Applied Information Technology 31^st July 2022 -- Vol. 100. No. 14 -- 2022
Full Text

Title:	DEEP LEARNING BASED METHOD FOR THE ACTIVITY OF NOBEL CORONAVIRUS DISEASE PREDICTION FROM THE MEDICAL RADIOGRAPH CHEST X-RAY IMAGES
Author:	USMAN HARUNA, ROZNIZA ALI, MUSTAFA MAN
Abstract:	Nearly 2 years ago, a new virus called Severe Acute Respiratory Syndrome CoronaVirus 2 (SARS-CoV-2) which is often referred to as Covid-19 was declared to be a pandemic by “World Health Organization” (WHO). The virus is among the most deadly virus diseases in the world, which has a high percentage of mortality and widespread rates. The standard procedure used for diagnosing suspected Coronavirus patients is through the use of kits called Real-time Reverse Transcriptase-Polymerase Chain Reaction (RT-PCR) kits that are usually available in small scale quantities and in addition to their high demand to spend a significant amount of time in the laboratory in the process of determining whether the suspected patients are Coronavirus positive or negative, which might result in increasing the chances to spread the virus. Radiologists discovered the presence of changes in radiograph Chest X-Ray medical images and Computer Tomography images whose ability is to detect the presence of the Coronavirus disease with higher sensitivity than using the RT-PCR, where the X-Ray images happened to be the most affordable among the duo. Therefore, there is a need to provide a kind of rapid diagnostic alternative that can be used to detect the presence of Coronavirus, “Covid-19”, that will control its spreading. Several radiological Chest X-Ray radiograph images have been used to propose a model that is capable of diagnosing Covid-19 patients through applying Convolutional neural networks (CNN). Two fine-tuned pre-trained models are compared to determine the model with the best performance. The experimental results show that the model using fine_tuned pre-trained VGG16 has achieved 92.50% of classification accuracy with 93.89% of sensitivity, 91.11% of specificity, 91.35% of precision, and 92.60% of F1_Score; while the results of the proposed model using fine-tuned pre-trained MobileNet has achieved 98.82% of classification accuracy with 100% of sensitivity, 97.64% of Specificity, 97.69% of Precision, and 98.83% of F1-Score. This revealed that MobileNet outperformed the VGG16 in the classification of the Chest X-Ray radiograph images for Coronavirus detection. It further indicates that even without applying many pre-processing techniques as did in creating existing models, our proposed model can perform better than many of them. This system can also be used in a situation where experts and clinical test kits are insufficient. The proposed model can also be used to fast track the time required to detect the Coronavirus disease by applying Chest X-Ray radiograph images of the suspected patients.
Keywords:	CNN, Covid-19, Fine_tuned, MobileNet, VGG16
Source:	Journal of Theoretical and Applied Information Technology 31^st July 2022 -- Vol. 100. No. 14 -- 2022
Full Text