Univrsity Of Tehran PressJournal of Information Technology Management2980-797218320260701RES2NET-MAXOUT: Image-Based Multi-Class Malware Classification with Deep Learning Models12410800210.22059/jitm.2026.388431.3967ENG.Joel SunnyDeolAssociate prof., Kallam HaranadhaReddy Institute of Technology (KHIT) (AUTONOMOUS), Chowdavaram, Andhra Pradesh, India.0009-0002-1018-3325K. Siva RamaPrasadAssistant Prof., Department of Computer and Engineering- Artificial intelligence & Machine learning, Kallam HaranadhaReddy Institute of Technology (KHIT) (AUTONOMOUS), Chowdavaram, Andhra Pradesh, India.A. HanumatPrasadAssociate Prof., Department of Computer and Engineering- Artificial intelligence & Machine learning, Kallam Haranadhareddy Institute of Technology (KHIT) (AUTONOMOUS), Chowdavaram, Andhra Pradesh, India.M. ChennakesavaRaoAssociate prof., Computer and Engineering- Artificial intelligence & Machine learning, Kallam HaranadhaReddy Institute of Technology (KHIT) (AUTONOMOUS), Chowdavaram, Andhra Pradesh, India.MedidaAlekhyaAssistant prof., Computer and Engineering- Artificial intelligence & Machine learning, Kallam HaranadhaReddy Institute of Technology (KHIT) (AUTONOMOUS), Chowdavaram, Andhra Pradesh, India.T.AlekhyaAssistant prof., Computer and Engineering- Artificial intelligence & Machine learning, Kallam HaranadhaReddy Institute of Technology (KHIT) (AUTONOMOUS), Chowdavaram, Andhra Pradesh, India.Journal Article20250108The dynamic nature of the Internet has resulted in an increase in network security threats, primarily caused by the growing incidence of malware and the continuous emergence of its many forms. Conventional methods for detecting malware involve complex feature engineering, which considerably reduces detection effectiveness. Furthermore, machine learning algorithms, which were once a mainstay in the fight against cyber threats, are no longer able to accurately detect every new and complex malware variation. In contrast, deep learning approaches have the potential to be transformative. Their inherent flexibility and capacity for pattern recognition present a viable solution to the ever changing challenge of detecting a wide range of malware variations more reliably and effectively. For this reason, a unique deep learning based architecture is developed in this study to categorize malware types and protect networks. Three components comprise the most significant parts of this work: feature extraction, classification, and malware visualization. The suggested method first converts malware binary data into two dimensional images. Secondly, we introduce an automatic feature extractor that utilizes the Res2Net model to extract shared patterns among family members from the visualized images. Lastly, we propose a Deep Maxout Network designed to categorize malware using the extracted attributes. Moreover, data augmentation based on Conditional Generative Adversarial Networks (CGAN) addresses the issue of imbalanced data. The proposed technique is evaluated using the MaleVis, Microsoft Malware Classification Challenge (MMCC), and Drebin datasets. The experimental findings demonstrate that the proposed method can successfully and accurately classify malware compared to current state of the art techniques.https://jitm.ut.ac.ir/article_108002_3f8c855e1ccc3466ab76aac3acdd6556.pdfUnivrsity Of Tehran PressJournal of Information Technology Management2980-797218320260701A Rough Set–Based Process Mining Approach for Knowledge Intensive Processes: Insights from a New Product Development Case Study254610800310.22059/jitm.2026.402437.4270ENMehriChehrehpakPh.D. Candidate, Faculty of Management and Economy, Science and Research Branch, Islamic Azad University, Tehran, Iran.0009-0003-4400-5292AbbasToloeiProfessor, Faculty of Management and Economy, Science and Research Branch, Islamic Azad University, Tehran, Iran.0009-0000-3595-2659KamranMohammadkhaniProfessor, Faculty of Management and Economy, Science and Research Branch, Islamic Azad University, Tehran, Iran.0000-0001-5891-8381Journal Article20250915Knowledge-intensive processes (KiPs) have gained significant attention in modern organizations, where decision-making plays a fundamental role in the value chain. Therefore, the analysis of decision-making logs and the identification of decision rules and models are crucial. This paper presents a decision model for optimizing the New Product Development (NPD) process based on a case study at a prominent information technology company in Iran. The study leverages rough set theory and a fast reduction algorithm to analyze decision points within the NPD process. The model introduces a step-by-step methodology for identifying decision models at critical decision points, aiming to streamline decision-making and enhance process efficiency. The paper discusses the identification of critical decision features and the creation of decision trees for each decision point, thereby illuminating the impact of the decision model. Moreover, the study reports a reduction in the number of decisions and improved process efficiency, with a substantial 90% reduction recorded after implementing the model in a real-world scenario within the organization. The paper also highlights certain limitations and proposes future research directions, emphasizing the potential for applying the model to other knowledge-intensive processes and within larger organizational contexts. The comprehensive review of the case study and the proposed decision model offer valuable insights into enhancing decision-making processes within the new product development domain.https://jitm.ut.ac.ir/article_108003_78b6b6f10fda9856f83266e716890126.pdfUnivrsity Of Tehran PressJournal of Information Technology Management2980-797218320260701Fusion of Graph-Based Ensembles Using Fuzzy Integrals for Persian Learning to Rank478110800410.22059/jitm.2026.412047.4443ENMaryamPiroozmandDepartment of Algorithms and Computation, School of Engineering Sciences, College of Engineering, University of Tehran, Tehran, Iran.0000-0001-9767-3911AliMoeiniProfessor, Department of Algorithms and Computation, School of Engineering Sciences, College of Engineering, University of Tehran, Tehran, Iran.0000-0002-6408-3525MojtabaMazoochiAssistant prof., ICT Research Institute, Tehran, Iran.0000-0001-9289-7940Amir HoseinKeyhanipourAssistant prof., Computer Engineering Department, Faculty of Engineering, College of Farabi, University of Tehran, Tehran, Iran.0000-0003-4137-9494Journal Article20260226The rapid growth of Persian-language Web content has created an urgent need for effective Learning to Rank (LTR) systems. However, developing such systems for Persian presents significant challenges, including Persian’s morphological complexity, its unique Web link topology, and the scarcity of annotated resources relative to dominant languages such as English. Existing methods often rely on hand-crafted features, apply graph techniques in isolation, or employ simple ensemble fusion, lacking a unified framework to integrate structural semantics and handle model uncertainty. This paper proposes PersianRank, a novel LTR framework that bridges these gaps by modeling query-document relationships as a weighted bipartite graph to extract graph-based structural features. It applies deep neural networks to predict these graph features for unseen query-document pairs, enabling feature augmentation during the testing phase. A diverse ensemble of base rankers is then trained on the enriched feature set, and their predictions are fused via the Choquet fuzzy integral, which performs a non-linear, uncertainty-aware aggregation. Evaluated on the dotIR dataset, PersianRank achieves state-of-the-art performance, improving P@1 by 17.1% relative to the best baseline (MDPRank) and demonstrating notable top-rank accuracy through effective graph-based feature space expansion and uncertainty-aware model fusion. The framework is especially effective for top-rank retrieval, a critical aspect in user-centric search systems.https://jitm.ut.ac.ir/article_108004_ee8960502935cdee25c4e36fd2808a56.pdfUnivrsity Of Tehran PressJournal of Information Technology Management2980-797218320260701Risk-Driven Capability Development in Global Supply Chain Resilience: Implications for Digitalization8210710800510.22059/jitm.2026.414800.4475ENHadiZareaDepartment of Management, Faculty of Business Administration, Laval University, Quebec, Canada.0000-0001-8970-6744HadiMoghaddamzadehAssistant Prof., Human resource management, Aras campus, University of Tehran, Tehran, Iran.0000-0001-8970-6744MahmoudMoradiAssociate Prof., Department of Industrial Management, Faculty of Management and Economics, University of Guilan. Rasht, Iran.AbdeslamHassaniProf., Department of Marketing and Information Systems, Business School, Université du Québec à Trois-Rivières, Quebec, Canada.ThierryWarinProfessor of Data Science for International Business, Department of International Business, HEC Montreal, Montreal, Canada.AleksanderStojkovProfessor of Economics, Department of Business Law and Economics, Ss. Cyril and Methodius University in Skopje. N. Macedonia.Journal Article20260519This study examines supply chain risk management and resilience capabilities in the automotive parts manufacturing sector, with particular attention to the role of digitalization in mitigating high-priority risks. Using a four-phase methodology, the research first identifies 35 supply chain risks from the literature, categorizing them into supply, operational, demand, and environmental risks. In the second phase, a survey of 164 managers from 47 firms refines these risks into 18 key factors using Exploratory Factor Analysis (EFA) and Confirmatory Factor Analysis (CFA). The third phase employs the Fuzzy Analytic Hierarchy Process (Fuzzy AHP), integrating pairwise comparisons from 21 CEOs to assess the probability and impact of each risk, classifying them into three action groups: immediate action, observation, and no action. Seven risks fall into the immediate action group: quality risk, supplier capacity disruption, sourcing flexibility risk, resource risk, market risk and competition, geopolitical risk, and macroeconomic and technological risk. In the final phase, each immediate action risk is mapped to specific firm capabilities, with emphasis on how digital technologies, including supply chain visibility platforms, data analytics, and digital procurement tools, can strengthen these capabilities and sustain firms within the zone of balanced resilience. The findings provide actionable guidance for managers on aligning digitalization investments with the most critical risk exposures, avoiding capability gaps and profit erosion. The study concludes by proposing a risk-driven capability development framework with implications for digital transformation in global supply chains.https://jitm.ut.ac.ir/article_108005_a47a2bf266954aaddd504ed4b3d1d8be.pdfUnivrsity Of Tehran PressJournal of Information Technology Management2980-797218320260701Bitcoin and Gold in Data-Driven Financial Risk Modelling: A Time-Varying Analytics Framework for Dual Financial Markets10813210800610.22059/jitm.2026.108006ENRajaRehanCollege of Business Management, Institute of Business Management, Pakistan.BahadurShah ZafarInstitute of Business Studies, Kohat University of Science and Technology, Pakistan.EhsanUllahGraduate School of Economics, Ural Federal University, Russia; Institute of Business Studies, Kohat University of Science and Technology, Pakistan.Gholam RezaZandiBusiness School, Universiti Kuala Lumpur, Malaysia.0000-0001-9517-8474Journal Article20260720This study examines the time-varying relationships and hedging properties of Bitcoin and gold with respect to Pakistani Islamic and conventional equity markets, the exchange rate, and inflation under conditions of economic turbulence. Using daily data from August 1, 2010, to March 1, 2024, the study employs a data-driven Dynamic Conditional Correlation GARCH (DCC-GARCH) framework combined with an extreme quantile approach to conduct the empirical analysis. The results show that gold serves as a strong hedge against both conventional and Islamic stocks, as well as against exchange rate and inflation. Moreover, gold has strong protective ability against declines in conventional stocks, exchange rate risk, and high inflation risk. Further, Bitcoin is found to act not only as a hedge against conventional stocks but also as a diversifier for the Islamic stock index. However, it fails to play a safe-haven role against both stock markets and inflation risk in turbulent conditions. These findings offer valuable insights for portfolio managers, policymakers, and investors seeking to enhance portfolio risk management during periods of economic uncertainty and market turbulence.https://jitm.ut.ac.ir/article_108006_6c7551777bbacc146e723831501d9e78.pdfUnivrsity Of Tehran PressJournal of Information Technology Management2980-797218320260701Nurses' Opportunities Regarding the Importance of Artificial Intelligence Application in Pediatric Wards13314610800710.22059/jitm.2026.108007ENLena HusainAbdu-AllahBabylon Health Directorate, Babylon, Iraq.0000-0002-7641-992XKamesBanderProfessor, Child Health Nursing, College of Nursing, University of Babylon, Iraq.Journal Article20260720Artificial intelligence (AI) is increasingly used in healthcare to advance patient care and nursing practice, allowing nurses to focus more on direct patient care and complex decision-making. This descriptive, cross-sectional study aimed to assess nurses' opportunities regarding the importance of AI applications in pediatric wards and to identify relationships with socio-demographic data. Conducted in Hillah, Iraq, from September 2025 to June 2026 at three hospitals (AL-Imam Al-Sadiq Teaching Hospital, Babel Teaching for Child and Maternity Hospital, and AL Noor Hospital), the study used a convenience sample of 250 nurses. Data were collected via a questionnaire and analyzed using descriptive and inferential statistics. Results showed that 36% of participants perceived moderate opportunities for AI use (mean ± SD = 3.34 ± 0.75), and significant relationships were found between opportunities and education level, department unit, and source of AI information (p ≤ 0.05). The study concludes that one-third of nurses perceive moderate opportunities for AI in pediatric wards, and recommends that hospitals improve their technology infrastructure to enable safe and efficient AI integration.https://jitm.ut.ac.ir/article_108007_b3876fa947ecae4f921d4167f08d0f9c.pdfUnivrsity Of Tehran PressJournal of Information Technology Management2980-797218320260701Digital Mammogram Image Feature Extraction Using Local Binary Patterns14715410800810.22059/jitm.2026.108008ENMusbahJumah AqelDepartment of Computer Science, Faculty of Information Technology, Zarqa University, Jordan.Journal Article20260720Nowadays, breast cancer is a very common disease among women and causes many deaths world-wide. One of the most popular techniques to diagnose breast cancer is using computer-aided diagno-sis (CAD) techniques. It helps diagnose cancer in its early stages. It can show cancer in different forms, such as masses, density, or calcifications. It is important to note that breast density is not the same as density in normal cases. In mammograms, breast density is evaluated by the percentage of fatty tissue in the breast, which differs from glandular tissue. Moreover, calcifications and masses are also important for determining breast cancer, but breast density is mainly used to identify breast cancer risk. Feature extraction is very helpful and facilitates the classification process. An algorithm for mammogram feature extraction based on fuzzy clustering and histograms was proposed to en-hance the feature extraction process. The proposed algorithm was implemented using MATLAB. A set of mammogram images was selected from the MIAS database, and the features extracted for these selected images were successful. However, the same set of images was also processed using the K-means clustering algorithm to compare the obtained features.https://jitm.ut.ac.ir/article_108008_e8ccb0cfba905533bae37109e65ee5e2.pdfUnivrsity Of Tehran PressJournal of Information Technology Management2980-797218320260701Twitt Ray: Sentiment Visualization of Arabic Tweets15518010800910.22059/jitm.2026.108009ENNadiaAl-GhreimilAssistant Prof., Department of Information Technology, College of Computer and Information Sciences, King Saud University, Riyadh, Saudi Arabia.ManalAlhassounData Scientist and Research Associate, King Abdulaziz City for Science and Technology, Riyadh, Saudi Arabia.NoraAl-TwaireshAssociate Professor, Department of Information Technology, College of Computer and Information Sciences, King Saud University, Riyadh, Saudi Arabia.DuaaAlSaeedProf., College of Computer and Information Sciences, King Saud University, Riyadh, Saudi Arabia.Journal Article20260720Sentiment Analysis of Twitter offers valuable insight into public sentiment. However, humans absorb most information through visual perception. Therefore, visual representation of data enables them to comprehend data more quickly and effectively than in text form. Unfortunately, the majority of ex-isting work mainly focuses on sentiment analysis and pays less attention to sentiment visualization; this potentially reduces the real value of sentiment analysis results. In this paper, we present an easy-to-use sentiment visualization tool for Arabic tweets called Twitt Ray. The goal of Twitt Ray is to serve anyone, from individuals to companies and social scientists, who are interested in knowing, analyzing, and visualizing the current sentiments that people express on Twitter. Moreover, Twitt Ray provides seven different visual representations of sentimentally analyzed data to achieve a better understanding of different aspects of the data. Also, a usability study was performed to evaluate the usability of the tool for end users.https://jitm.ut.ac.ir/article_108009_1a18be5fc0a00b43a51b9fa3d99c401c.pdfUnivrsity Of Tehran PressJournal of Information Technology Management2980-797218320260701Generative AI Health Assistants: From Theory to Practice18120110801010.22059/jitm.2026.108010ENGhadaAlhudhudProf., Department of Data Science and Artificial Intelligence, Faculty of Information Technology, Al-Ahilya Amman University, Amman, Jordan.0000-0002-2810-6582SanaaAlhalabiAssistant Prof., Department of Computer Science, Faculty of Information Technology, Al-Ahilya Amman University, Amman, Jordan.0009-0002-2839-0349MarwanAl-AkaidiProfessor, Vice Chancellor and CEO Wigwe University, Isiokpo, Nigeria.Journal Article20260720Generative Artificial Intelligence (GenAI) has evolved significantly from early models like variational autoencoders (VAEs) to sophisticated diffusion and hybrid models, enabling the creation of entirely new data from input prompts. In the domain of natural language, powerful large language models (LLMs) such as GPT-4 now generate remarkably coherent responses, while diffusion models have opened up new applications in text-to-image generation and image editing across unimodal and multimodal spaces. These advancements are built upon foundational theories of latent variable modeling, denoising processes, and attention/transformer architectures. This paper traces the development of GenAI, with a focus on its application in mobile AI assistants integrated into smartphones and edge computing platforms. While transformative, these models are typically large and resource-intensive. Their deployment in mobile or edge environments necessitates architectural compression techniques like pruning and quantization, model partitioning between the device and server, or the development of lightweight models designed for constrained settings. This study discusses the practical transition from model architecture to a functional, AI-powered mobile assistant that utilizes dynamic inference across device and edge layers. A mixed-methods approach is employed, combining quantitative benchmarking of latency, accuracy, energy consumption, and resource utilization with qualitative assessments of user experience, trust, and privacy perceptions. The system's design incorporates principles of data protection and user authentication through voice biometric-based watermarking, aligning with our previous work on secure multimedia exchange. Using a heterogeneous dataset of benchmarked prompts, mobile hardware telemetry, and semi-structured user interviews, this research reports on performance metrics, usability findings, and privacy trade-offs. The paper concludes by offering recommended practices and reference architectures, providing a holistic framework for transitioning GenAI from a theoretical innovation to a practical, secure, and ethically aligned tool in mobile and health-oriented applications.https://jitm.ut.ac.ir/article_108010_fe78e3a154c6b0b2934e16b54a315bdc.pdfUnivrsity Of Tehran PressJournal of Information Technology Management2980-797218320260701Advancing Automatic Sarcasm Detection in the French Language20222710801110.22059/jitm.2026.414898.4482ENYannick U.Tchantchou SamenAssistant Prof., Senior Lecturer, Ph.D., Department of Mathematics and Computer Science, Faculty of Science, University of Maroua, P.O Box:814 Maroua, Cameroon; Laboratoire de Recherche en Sciences Informatiques et Applications (LRSIA), UAC, Abomey-Calavi, Benin.0000-0002-4031-4317Abou A.DjafarPh.D. Candidate, Department of Mathematics and computer Science, Faculty of Science, University of Maroua, Maroua, Cameroon.KaladzaviGuidediAssociate Prof., Department of Computer Science and telecommunications, National Advanced School of Engineering of Maroua, University of Maroua, Maroua, Cameroon.Fréjus A. A.LaleyeOpscidia, 9 Rue des colonnes, Paris, France.Journal Article20260522The proliferation of user-generated content necessitates robust sentiment analysis and opinion mining. Sarcasm detection presents a significant challenge in natural language processing, as sarcastic statements often invert literal meaning. Accurate sarcasm detection is crucial for enhancing sentiment analysis, opinion mining, recommender systems, and public opinion monitoring. While transformer-based and deep learning models have advanced sarcasm detection in English and other languages, French remains underdeveloped due to limited annotated datasets and the absence of models tailored to its linguistic nuances. Consequently, existing methods struggle to generalize and fail to leverage complementary semantic information to improve sarcasm recognition. This paper introduces a novel hybrid multi-task learning framework for French sarcasm detection. This architecture integrates transfer learning with sentiment analysis by combining a fine-tuned CamemBERT encoder, an auxiliary sentiment-analysis module, and a Bi-LSTM classifier. This approach jointly models contextual and affective information to better distinguish between sarcastic and non-sarcastic text. A significant contribution of this study is the development and release of a large-scale French sarcasm dataset comprising approximately 22,000 manually collected and curated annotated instances from diverse sources, providing a valuable benchmark for future research. Another key contribution is the proposed novel hybrid multi-task architecture that integrates transfer learning, sentiment analysis, and Bi-LSTM-based classification for improved sarcasm detection. Empirical evidence shows that sentiment-aware representations significantly improve sarcasm detection over standard transformer approaches. Experiments on the TransCasm dataset and the new corpus demonstrate the model's effectiveness, achieving F1-scores of 96% and 90%, respectively, surpassing prior results.https://jitm.ut.ac.ir/article_108011_cb19319caeb1084ebe48cc34834233ff.pdf