Bekim Fetaji

Dr. Bekim Fetaji

orcid

Research Focus & Metrics

Academic Career Length

9Years (2018-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
3 March 2025 in IEEE Access
Verified
Journal Article
Open access
BFBekim Fetaji
Bekim Fetaji
MFMajlinda Fetaji
Majlinda Fetaji
AHAffan Hasan
Affan Hasan
SRShpetim Rexhepi
Shpetim Rexhepi
GAGoce Armenski
Goce Armenski
Online financial fraud remains a pervasive threat, incurring billions of dollars in global losses annually. Mid-sized markets, such as North Macedonia, face acute challenges as digital adoption in the Banking, Financial Services, and Insurance (BFSI) sector outpaces the establishment of robust, multi-layered security systems. This paper introduces FRAUD-X, a unified framework merging artificial intelligence (AI)–based anomaly detection, blockchain-driven transaction verification, cybersecurity intrusion detection, and real-time early warning mechanisms into a single pipeline. Drawing upon three datasets—a Credit Card Fraud dataset (Kaggle), the PaySim Mobile Money dataset, and collected 50,000 anonymized local BFSI transactions from North Macedonia—FRAUD-X demonstrates a ~2–4% improvement in F1 compared to single-plane AI approaches, with ~90% recall for zero-day threats. Key enhancements include: (1) a permissioned blockchain for tamper-proof ledger entries, (2) synergistic AI–cybersecurity integration for dynamic risk scoring, and (3) real-time alerts that reduce reaction windows from hours to mere minutes. The framework runs at ~15–16 ms per transaction (~33% CPU usage), supporting near-real-time BFSI operations. Ablation studies confirm that each synergy layer (blockchain, cybersecurity, and early warning) significantly contributes to overall performance. A security analysis illustrates how FRAUD-X mitigates node compromise, collusion attempts, and advanced persistent threats (APT). By providing a replicable roadmap that balances high detection accuracy with operational feasibility, FRAUD-X offers practical value to BFSI entities in North Macedonia and comparable mid-scale markets.

2 September 2025 in Mesopotamian Journal of Computer Science
Verified
Journal Article
Open access
PDF file
BFBekim Fetaji
Bekim Fetaji
DEBABRATA SAMANTADEBABRATA SAMANTA
DEBABRATA SAMANTA
The paper fills in a gap in the literature that demonstrates an insufficient number of sturdy detection schemes that can recognize the small semantic aberrations inherent in LLM-generated deceptive text. Our proposed co-design hybrid model is Semantic Anomaly Detection with Isolation Forest (SADI) model that combines the synergistic mixture of a fine-tuned transformer-based LLM for deep semantic feature extraction with Isolation Forest algorithm that detects anomalies efficiently. This study introduces SADI, an adaptive semantic-anomaly detector for large-language-model phishing emails. Using a corpus of 10 000 messages, SADI attains an F1 score of 0.981 (95 % CI 0.978–0.984) and processes a single message in 18 ms on consumer GPUs. An expanded evaluation against three public benchmarks and a live enterprise feed confirms robustness to prompt variation. Code, data splits, and a reproducible environment file accompany the paper. We also prepared the new, more challenging target dataset, that is phishing attacks synthesized by a variety of state-of-the-art LLMs, denoted LLM-Phish-Synth-2025, with this objective in mind. The results of our experiments on three publicly available data sets and our new corpus show that SADI received a higher F1- score of 0.981 compared to the baseline models, including separate fine-tuned LLMs, by a wide margin. The proposed SADI architecture is the first to combine semantic anomaly detection with adaptive, contamination-aware isolation in the context of LLM-generated phishing, addressing both scalability and evolving attack sophistication. The theoretical impact is the new architecture of architectural fusion that branched semantic anomaly detection and the practical advantage of a more robust defense solution against an ever-evolving threat in cyber space in addition to the provision of a new benchmark dataset to the research community. This approach is an efficient and scalable solution to combat the wave of the phishing campaigns generated by AI.

1 December 2020 in New Trends and Issues Proceedings on Humanities and Social Sciences
Verified
Journal Article
Open access
PDF file
BFBekim Fetaji
Bekim Fetaji
MFMajlinda Fetaji
Majlinda Fetaji
OAOzcan Asilkan
Ozcan Asilkan
MEMirlinda Ebibi
Mirlinda Ebibi
In the last decade, applications of virtual reality (VR) and augmented reality (AR) technologies have rapidly increased. Although they are used especially in the game industry, their importance in education has also becoming more evident. As there are not enough academic researches on this field, we aimed to examine the role of VR and AR technologies in education. During the study, the implementation of VR and AR applications in education was analysed as well as impact factors, barriers and levels of knowledge transfer were investigated. We propose an approach to address the various identified issues. This study aims to contribute to the field with the comparison of the results, followed by analyses and recommendations for future work. Keywords: Virtual reality, augmented reality, education, learning models, assessment, flipped learning.

4 November 2021
Verified
Book Chapter
Closed access
PDF file
BFBekim Fetaji
Bekim Fetaji
MFMajlinda Fetaji
Majlinda Fetaji
MEMirlinda Ebibi
Mirlinda Ebibi
MAMaaruf Ali
Maaruf Ali

27 November 2024 in TEM Journal
Verified
Journal Article
Open access
BFBekim Fetaji
Bekim Fetaji
MFMajlinda Fetaji
Majlinda Fetaji
MEMirlinda Ebibi
Mirlinda Ebibi
FFFjolla Fetaji
Fjolla Fetaji
Traditional instructional approaches often fail to engage students and solidify conceptual understanding. This study proposes the VIMATE (Visualization Impact on Mathematical Transformations Assessment in Education) framework and introduces a novel empirical approach for evaluating the role of visualization tools in enhancing student comprehension of geometric transformations. In order to conduct this analysis, an original dataset was obtained from 300 high school students. The VIMATE framework involves several components of the research and is complemented by a literature review and meta-analysis of selected mathematical software platforms, such as MATLAB, Mathematica and GeoGebra, and quasi-experimental student assessments, as well as, questionnaires capturing the teachers’ and learners’ opinions. The overall aim of this approach is to fill major gaps in the prior research by offering an integrated methodology to objectively evaluate these visualization instruments on student achievement, reasoning abilities, and problem-solving strategies. The research makes four pivotal contributions by presenting the VIMATE framework as the methodological approach for assessing the use of educational technology in teaching practice.