Explainable Anomaly Detection for Digital Crowdsourcing Platforms Using Optimized Isolation Forest and Local Outlier Factor

MOHD NORHISHAM RAZALI; Norizuandi Ibrahim; Sokkhey PHAUK; Rozita Hanapi; Norlelawati Ismawi

doi:10.35877/454RI.asci4873

Authors

MOHD NORHISHAM RAZALI UNIVERSITI TEKNOLOGI MARA
Norizuandi Ibrahim
Sokkhey PHAUK
Rozita Hanapi
Norlelawati Ismawi

DOI:

https://doi.org/10.35877/454RI.asci4873

Keywords:

Anomaly Detection, Hyperparameter Optimization, Explainable AI

Abstract

Digital crowdsourcing platforms depend on freelancer profiles to support trust, reputation, and hiring decisions. However, anomalous profiles with unusual attribute combinations may reduce platform reliability. Existing anomaly detection studies mainly emphasize detection accuracy while overlooking robustness, interpretability, and governance relevance. This study proposes an explainable anomaly detection framework for identifying anomalous freelancer profiles using Isolation Forest (IF) and Local Outlier Factor (LOF). Hyperparameter optimization techniques, including Grid Search, Randomized Search, Genetic Algorithms, and Bayesian Optimization, are applied to improve model robustness. Beyond detection performance, the framework evaluates anomaly consistency, clustering separability, and cross-model agreement. To improve transparency, SHAP-based explainability is integrated to interpret anomaly scores and identify feature-level contributions. This enables platform administrators to understand why freelancer profiles are classified as anomalous and supports human-in-the-loop decision-making. Experimental results show that optimized models consistently detect economically and reputationally implausible profiles while reducing unstable detections. The findings demonstrate that integrating optimization and explainable AI enhances trustworthy anomaly detection and supports governance in digital crowdsourcing platforms.

Downloads

Download data is not yet available.

References

Agyemang, E. F. (2024). Anomaly detection using unsupervised machine learning algorithms: A simulation study. Scientific African, 26. https://doi.org/10.1016/j.sciaf.2024.e02386

Bobek, S., Kuk, M., Szelazek, M., & Nalepa, G. J. (2022). Enhancing Cluster Analysis With Explainable AI and Multidimensional Cluster Prototypes. IEEE Access, 10, 101556 – 101574. https://doi.org/10.1109/ACCESS.2022.3208957

Carletti, M., Terzi, M., & Susto, G. A. (2023). Interpretable Anomaly Detection with DIFFI: Depth-based Isolation Forest Feature Importance. http://arxiv.org/abs/2007.11117

Carletti, M., Terzi, M., & Susto, G. A. (2023). Interpretable Anomaly Detection with DIFFI: Depth-based feature importance of Isolation Forest. Engineering Applications of Artificial Intelligence, 119. https://doi.org/10.1016/j.engappai.2022.105730

Chaabouni, A., & Boujelben, M. A. (2025). Outlier Ensemble Based on Isolation Forest: The CBOEA Approach. Foundations of Computing and Decision Sciences, 50(1), 27 – 55. https://doi.org/10.2478/fcds-2025-0002

Darrab, S., Allipilli, H., Ghani, S., Changaramkulath, H., Koneru, S., Broneske, D., & Saake, G. (2024). Anomaly Detection Algorithms: Comparative Analysis and Explainability Perspectives. Communications in Computer and Information Science, 1943 CCIS, 90 – 104. https://doi.org/10.1007/978-981-99-8696-5_7

Dehkordi, S. B., Nasri, S., & Dami, S. (2025). Advances in network security and new anomaly detection techniques. Majlesi Journal of Electrical Engineering, 19(2). https://doi.org/10.57647/j.mjee.2025.1902.26

Hermosilla, P., Berríos, S., & Allende-Cid, H. (2025). Explainable AI for Forensic Analysis: A Comparative Study of SHAP and LIME in Intrusion Detection Models. Applied Sciences (Switzerland), 15(13). https://doi.org/10.3390/app15137329

Khalid, Z., Iqbal, F., & Saqib, M. (2025). Bridging knowledge gaps in digital forensics using unsupervised explainable AI. Forensic Science International: Digital Investigation, 53. https://doi.org/10.1016/j.fsidi.2025.301924

Kiefer, S., & Pesch, G. (2021). Unsupervised Anomaly Detection for Financial Auditing with Model-Agnostic Explanations. Lecture Notes in Computer Science (Including Subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), 12873 LNAI, 291 – 308. https://doi.org/10.1007/978-3-030-87626-5_22

Li, X., Guo, H., Xu, L., & Xing, Z. (2023). Bayesian-Based Hyperparameter Optimization of 1D-CNN for Structural Anomaly Detection. Sensors, 23(11). https://doi.org/10.3390/s23115058

Lunawat, S., Rao, J., & Patil, P. (2024). A Comprehensive Survey on Anomaly Detection in Social Media Networks: Challenges, Methods, and Future Directions. 4th International Conference on Sustainable Expert Systems, ICSES 2024 - Proceedings, 363 – 370. https://doi.org/10.1109/ICSES63445.2024.10763303

Mohale, V. Z., & Obagbuwa, I. C. (2025). Evaluating machine learning-based intrusion detection systems with explainable AI: enhancing transparency and interpretability. Frontiers in Computer Science, 7. https://doi.org/10.3389/fcomp.2025.1520741

Mohanty, M., Rath, P. S., Mohapatra, A. G., Mohanty, A., & Senapati, S. K. (2025). AI-enhanced data processing for modeling applications. Advances in Computers. https://doi.org/10.1016/bs.adcom.2025.09.004

Morichetta, A., Casas, P., & Mellia, M. (2019). Explain-IT: Towards explainable AI for unsupervised network traffic analysis. Big-DAMA 2019 - Proceedings of the 3rd ACM CoNEXT Workshop on Big DAta, Machine Learning and Artificial Intelligence for Data Communication Networks, Part of CoNEXT 2019, 22 – 28. https://doi.org/10.1145/3359992.3366639

Sakthivanitha, M., Priscila, S. S., & Praveen, B. M. (2025). E-commerce Security: An Overview of AI Driven Threat / Anomaly Detection. International Journal of Engineering Trends and Technology, 73(6), 157 – 172. https://doi.org/10.14445/22315381/IJETT-V73I6P114

Vajpayee, P., & Hossain, G. (2024). Reduction of Cyber Value at Risk (CVaR) Through AI Enabled Anomaly Detection. Conference Proceedings - IEEE SOUTHEASTCON, 623 – 629. https://doi.org/10.1109/SoutheastCon52093.2024.10500040

Wang, S., Zhang, L., & Huang, X. (2026). Indoor WiFi Localization via Robust Fingerprint Reconstruction and Multi-Mechanism Adaptive PSO-LSSVM Optimization. Applied Sciences (Switzerland), 16(2). https://doi.org/10.3390/app16020753

Witayanont, Y., & Viyanon, W. (2025). Anomaly Detection in Bitcoin Network: Using Distance-based and Tree-based Unsupervised Learning Methods. Proceedings of the 6th ACM International Symposium on Blockchain and Secure Critical Infrastructure, BSCI 2024. https://doi.org/10.1145/3659463.3660022

Yahia, A., Mouhssine, Y., Elalaoui, A., & Ouatik Elalaoui, S. (2024). Leveraging Machine Learning for Anomaly Detection Methods in Cryptocurrency: A Data-Driven Study. 10th Edition of the International Conference on Optimization and Applications, ICOA 2024 - Proceedings. https://doi.org/10.1109/ICOA62581.2024.10754457

Younis, S. B. (2026). An Immersive Digital Twin Framework for Intelligent Spatially Grounded Anomaly Detection in Smart Homes. International Journal of Intelligent Engineering and Systems, 19(2), 1016 – 1030. https://doi.org/10.22266/ijies2026.0228.61

Zhang, Y.-Q., Lv, S.-Q., & Fan, D. (2015). Anomaly detection in online social networks. Jisuanji Xuebao/Chinese Journal of Computers, 38(10), 2011 – 2027. https://doi.org/10.11897/SP.J.1016.2015.02011