An Interpretable Hybrid Model for Fraud and Threat Detection in E-commerce

DOI

10.2991/978-94-6239-616-6_24How to use a DOI?

Keywords

Explainable Artificial Intelligence (XAI); Random Forest; Local Interpretable Model-Agnostic Explanations; ECOMNET; CNN-LSTM; Grad-CAM

Abstract

With the rapid growth of e-commerce platforms, ensuring secure and trustworthy user interactions has become a major challenge. Anomalies such as fraudulent transactions, fake reviews, and bot-driven activities pose significant threats to platform integrity and customer trust. Traditional anomaly detection models often lack transparency, making it difficult for stakeholders to interpret decisions and take informed actions. This project proposes an Explainable Artificial Intelligence (XAI) framework to enhance threat management in e-commerce by integrating interpretability into anomaly detection models. We utilize Random Forest for structured data anomaly detection, coupled with Local Interpretable Model-Agnostic Explanations (LIME) and SHapley Additive exPlanations (SHAP) to provide local and global model interpretability, respectively. For unstructured data such as user behavior logs or transaction image data, we employ a deep learning model named ECOMNET (a hybrid CNN-LSTM architecture), leveraging Gradient-weighted Class Activation Mapping (Grad-CAM) to visualize and explain the model’s focus during threat detection. The proposed framework not only improves anomaly detection accuracy but also enables domain experts and analysts to understand the “why” behind each prediction. This interpretability supports better decision-making, compliance with regulations, and increased trust in AI-driven security systems. Experimental results on real-world e-commerce datasets demonstrate that the integrated XAI techniques effectively balance performance with transparency, making the system robust, explainable, and actionable.

Copyright

© 2026 The Author(s)

Open Access

Open Access This chapter is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International License (http://creativecommons.org/licenses/by-nc/4.0/), which permits any noncommercial use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

Download article (PDF)

TY  - CONF
AU  - E. Valarmathi
AU  - J. Ragul
AU  - M. Ganesh Kumar
AU  - M. Surya
PY  - 2026
DA  - 2026/03/31
TI  - An Interpretable Hybrid Model for Fraud and Threat Detection in E-commerce
BT  - Proceedings of the International Conference on Artificial Intelligence and Secure Data Analytics (ICAISDA 2025)
PB  - Atlantis Press
SP  - 286
EP  - 298
SN  - 1951-6851
UR  - https://doi.org/10.2991/978-94-6239-616-6_24
DO  - 10.2991/978-94-6239-616-6_24
ID  - Valarmathi2026
ER  -