29 April 2026 in Bangladesh Journal of Multidisciplinary Scientific Research
MOMarwan Alshar’e Associate Professor, Faculty of Computing and IT, Sohar University, Sohar, Sultanate of Oman
Marwan Alshar’e Associate Professor, Faculty of Computing and IT, Sohar University, Sohar, Sultanate of Oman
RJRifat O. Shannak Professor, Business School, Jordan University, Amman, Jordan
Rifat O. Shannak Professor, Business School, Jordan University, Amman, Jordan
NONour Eldin Elshaiekh Osman Assistant Professor, Sultan Qaboos University, College of Arts and Social Sciences, Department of Information Studies, Sultanate of Oman
Nour Eldin Elshaiekh Osman Assistant Professor, Sultan Qaboos University, College of Arts and Social Sciences, Department of Information Studies, Sultanate of Oman
BJBasel Bani-Ismail Assistant Professor, Department of Software Engineering, Faculty of Information Technology, Applied Science Private University, Amman, Jordan
Basel Bani-Ismail Assistant Professor, Department of Software Engineering, Faculty of Information Technology, Applied Science Private University, Amman, Jordan
In financial cybersecurity, fraud detection is at the forefront of the battle against cybercriminals and deep learning models are the state-of-the-art method for real-time detection; yet their reliance on fixed feature distributions renders them inherently susceptible to adversarial attacks. With the rise of generative artificial intelligence (AI) in obfuscating fraudulent transactions, conventional performance indicators (accuracy and F1-score) do not reflect the cybersecurity resilience of these models. This research explores the Credit Card Fraud Detection dataset, comprising more than 550,000 anonymized records. The study uses a functional deep learning model as a benchmark and introduces the Fast Gradient Method (FGM) to simulate white box attacks on the latent features . The experiment results demonstrate a high detection accuracy of 99.85% under normal operating conditions, which dramatically reduces to 92.75% when card transactions are slightly modified by adversarial disturbances at an epsilon value of 0.1. Numerical evidence reveals a clear 7.10% "Security Gap," depicting a substantial degree of vulnerability, where transactions are mathematically reclassified from fraudulent to legitimate. Notably, our research suggests a non-linear "safety cliff" in the decrease of detection accuracy, where the model's reliability completely fails as the adversarial strength reaches 0.15 or higher. In addition, Principal Component Analysis demonstrates that gradient attacks have successfully distorted the latent features of fraudulent transactions, resulting in the movement of highly suspicious transactions towards the concentrated area of legitimate transactions, hence evading automated security screening while preserving the overall statistical distribution of the data.
1 August 2026 in Emerging Science Journal
Security and privacy (S&P) concerns are generally regarded as a barrier to students' competency transformation in higher education mobile e-learning. Yet others have not explored their impact. This work employs a cross-sectional quantitative design and analysis using ‘Partial Least Squares Structural Equation Modeling (PLS-SEM)’ to test the new model, the Security Competency Acceptance model (SCAM). Survey data were collected from 403 university students. The results show that mobile utility significantly impacts competency transformation (β = 0.370, p = 0.001) and soft skills (β = 0.355, p = 0.001). S&P issues significantly mediate the relationship between mobile utility and competency transformation (β= 0.223, p < 0.001). In addition, there was a significant sequential mediation effect through soft skills (β = 0.065, p = 0.001), showing that soft skills are the key element between the two variables. The model accounts for 50.3%, 46.5%, and 40.7% of the variance in S&P concerns, soft skills, and competency transformation, respectively. This research goes beyond conceptualizing S&P as simply a suppressive factor to treating it as a process that can facilitate learning, growth, and competence. In practical terms, it provides evidence that embedding privacy awareness in mobile learning is an intrinsic part of soft and graduate skills development, with implications for the design of a secure digital curriculum.
28 April 2026 in Bangladesh Journal of Multidisciplinary Scientific Research
MIMaruf Fatima Sadriwala Assistant Professor, Pacific University, Faculty of Commerce and Management, India
Maruf Fatima Sadriwala Assistant Professor, Pacific University, Faculty of Commerce and Management, India
RJRifat O. Shannak Professor, Business School, Jordan University, Amman, Jordan
Rifat O. Shannak Professor, Business School, Jordan University, Amman, Jordan
MOMarwan Alshar’e Associate Professor, Faculty of Computing and IT, Sohar University, Sohar, Sultanate of Oman
Marwan Alshar’e Associate Professor, Faculty of Computing and IT, Sohar University, Sohar, Sultanate of Oman
Existing business intelligence (BI) systems are predominantly descriptive and offer limited support for actionable decision-making in sales optimization, leading to a disconnect between data analytics and operational management. This study examines the development of a unified BI simulator that integrates machine learning, Pareto-based diagnostics, and what-if simulation to support data-driven sales optimization. The research employs an empirical pizza sales dataset from Kaggle, consisting of transactional, temporal, and operational variables. It applies multiple machine learning algorithms for comparative evaluation, along with Pareto analysis to identify key product contributors and a simulation engine to assess the sensitivity of low-performing products to operational changes. The results reveal that the Random Forest model outperforms other models, achieving 97.5% accuracy, an F1-score of 0.941, and an AUC of 0.996. Pareto analysis shows that approximately 30% of product categories account for nearly 80% of total sales, while a small proportion of products consistently exhibit low demand. Additionally, simulation analysis indicates that variations in operational factors, such as delivery efficiency, delivery distance, and product complexity, result in sales variability of up to 12.8%. The findings of this study suggest that the integration of predictive, diagnostic, and simulation-based analytics within a unified BI framework enables precise identification of key sales drivers and quantitatively measures the responsiveness of low-performing products to operational changes, thereby offering a comprehensive and data-driven evaluation of sales performance.