Attention-Guided Deep Representation Learning for Suspicious Activity and Anomaly Detection in Cybercrime Forensics Analysis
International Journal of Medical Toxicology and Forensic Medicine,
Vol. 16 (2026),
1 January 2026
https://doi.org/10.22037/ijmtfm.v16.52670
Abstract
Background: Cybercrime encompasses fraudulent activities conducted in virtual environments, such as identity theft, hacking, ransomware, and phishing attacks. Conventional cybercrime forensic analysis approaches are mainly based on statistical and rule-based approaches. Although recent machine learning and deep learning techniques have improved automated detection, existing approaches still suffer from repetitive features, insufficient dependency modeling, and limited detection performance in dynamic cyber environments.
Methods: This study proposes an Attention-Guided Deep Representation Learning for Effective Cybercrime Forensic Analysis (AGDRL-ECFA) framework. The contribution of the study lies in the design of an attention-guided deep representation-learning framework that improves anomaly detection accuracy and forensic pattern discrimination in large-scale cybercrime environments. For feature selection, the proposed model employs a modified ReliefF that identifies the most significant features and eliminates redundant ones. In addition, the attention-based conditional variational autoencoder is utilized to capture underlying data dependencies and classify anomalous or suspicious activities. Eventually, RMSProp is applied to improve training by dynamically adjusting the learning rate for each parameter.
Results: Simulation experiments are performed to ensure an enhanced outcome of AGDRL-ECFA system on the Cybercrime Forensic Dataset. The proposed model achieves better performance, with an accuracy of 97.73% compared to existing models.
Conclusion: The comparative result analysis demonstrates the improvement of the AGDRL-ECFA method and confirms its effectiveness for intelligent cybercrime forensic investigation and decision-support applications.
- Cybercrime Forensic Analysis
- Cyber Activities
- Deep Learning
- Digital Platforms
- Network Traffic
- Federated Learning
How to Cite
References
[1] Khan A. AI-driven cybercrime forensics for predictive threat detection and investigative intelligence. Int J Sci Interdiscip Res. 2026;7(1):280-324. [DOI: 10.63125/mggyg703]
[2] Islam U, Alwageed HS, Farooq MMU, Khan I, Awwad FA, Ali I, et al. Investigating the effectiveness of novel support vector neural network for anomaly detection in digital forensics data. Sensors. 2023;23(12):5626. [DOI: 10.3390/s23125626]
[3] Alsubaei FS, Almazroi AA, Ayub N. Enhancing phishing detection: a novel hybrid deep learning framework for cybercrime forensics. IEEE Access. 2024;12:8373-89. [DOI: 10.1109/ACCESS.2024.3351946]
[4] Ananjana WP, Arambawela JS, Gonawala DSN, Rathnayaka RGH, Senarathne AN, Siriwardena SDN. Machine learning-based criminal behavior analysis for enhanced digital forensics. PLoS One. 2025;20(10):e0332802. [DOI: 10.1371/journal.pone.0332802]
[5] Mardodaj J, Fetaji B. Integrating advanced algorithms and machine learning for cybercrime investigations. In: Proceedings of the International Conference on Intelligence-Based Transformations of Technology and Business Trends. Cham: Springer Nature Switzerland; 2025. p. 662-72. [DOI: 10.1007/978-3-032-07373-0_50]
[6] Islam U, Alsadhan AA, Alwageed HS, Al-Atawi AA, Mehmood G, Ayadi M, et al. SentinelFusion based machine learning comprehensive approach for enhanced computer forensics. PeerJ Comput Sci. 2024;10:e2183. [DOI: 10.7717/peerj-cs.2183]
[7] Khazem MM, Al Okagaili SM, Jassim BM, Hasan MS, Ahmed SA, Khalaf IZ, Mohammed AA. Real-time detection of cybercrime and digital forensics automation using AI-based legal evidence collection and preservation mechanisms. In: 2025 3rd International Conference on Cyber Resilience (ICCR); 2025 Jul. p. 1-7. [DOI: 10.1109/ICCR67387.2025.11292443]
[8] Puchalski D, Pawlicki M, Kozik R, Renk R, Choraś M. Trustworthy AI-based cyber-attack detector for network cyber crime forensics. In: Proceedings of the 19th International Conference on Availability, Reliability and Security; 2024 Jul. p. 1-8. [DOI: 10.1145/3664476.3670880]
[9] Kamble D, Bartere M. FrEnDTM: digital forensic investigation using frequency-enhanced attention-based deep learning model. SN Comput Sci. 2026;7(2):214. [DOI: 10.1007/s42979-026-04803-1]
[10] Babu MM, Harshini MS, Rao PK, Gowtham R, Jamuna G. Digital forensics for detecting and investigating cyber malicious activities. Int J Hum Comput Intell. 2026;5(3):744-55. [DOI: 10.5281/zenodo.18672461]
[11] Li GM, Liu N, Zhang JA. Speech emotion recognition based on modified ReliefF. Sensors. 2022;22(21):8152. [DOI: 10.3390/s22218152]
[12] Guo W, Sun S, Tang C, Li G, Bai X, Zhao Z. Classification of anomaly patterns in integrated energy systems based on conditional variational autoencoder and attention mechanism. Energies. 2023;16(11):4367. [DOI: 10.3390/en16114367]
[13] Jimohyusuf. Cybercrime forensic dataset [dataset]. Kaggle. [Link]
[14] Dunsin D, Ghanem MC, Ouazzane K, Vassilev V. A comprehensive analysis of the role of artificial intelligence and machine learning in modern digital forensics and incident response. Forensic Sci Int Digit Investig. 2024;48:301675. [DOI: 10.1016/j.fsidi.2023.301675]
[15] Ananjana WP, Arambawela JS, Gonawala DSN, Rathnayaka RGH, Senarathne AN, Siriwardena SDN. Machine learning-based criminal behavior analysis for enhanced digital forensics. PLoS One. 2025;20(10):e0332802. [DOI: 10.1371/journal.pone.0332802]
[16] Alsubaei FS, Almazroi AA, Ayub N. Enhancing phishing detection: a novel hybrid deep learning framework for cybercrime forensics. IEEE Access. 2024;12:8373-89. [DOI: 10.1109/ACCESS.2024.3351946]
- Abstract Viewed: 4 times
- PDF Downloaded: 1 times