Explainable Artificial Intelligence in Anomaly Detection for Threat Management in E- Commerce Platform
Summary
No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.
Summarise with VIRAT Free to try with your Science Catalogs account.Abstract
E-commerce platforms are increasingly exposed to diverse cyber threats such as fraudulent transactions, fake reviews, account takeovers, and abnormal user behaviors, all of which compromise security and diminish user confidence.Existing detection mechanisms, primarily based on ensemble methods such as Random Forest and XGBoost, often fail to achieve high predictive accuracy due to their limitations in capturing complex spatio-temporal dependencies.To overcome these challenges, this paper proposes a hybrid deep learning framework that integrates Convolutional Neural Networks (CNN) with Long Short-Term Memory (LSTM) networks.CNN is employed for internal spatial feature extraction, while LSTM is leveraged for modeling sequential and temporal dependencies, thus enabling the system to identify nuanced anomalies in user behavior and transaction patterns.The framework supports real-time threat monitoring, ensuring immediate detection of fraud, account takeovers, and abnormal behaviors, while providing interpretable reasoning for flagged anomalies.An interactive analyst dashboard is further introduced to visualize anomalies, risk scores, and feature importance, enabling human-in-the-loop feedback for continuous system refinement.By combining hybrid modeling, explainability, and analyst-driven feedback, the proposed approach significantly improves accuracy, reduces false positives, and strengthens trust in automated decision-making.Experimental validation demonstrates its potential to enhance resilience and operational security in modern e-commerce ecosystems.
Fields of Study & Subjects
Anomaly Detection Techniques and Applications
Imbalanced Data Classification Techniques
Network Security and Intrusion Detection
Publication details
- Journal / source
- Advances in intelligent systems research/Advances in Intelligent Systems Research
- Pages
- 1668–1678
- Publication date
- 1 April 2026
- Language
- EN
- Document type
- Book chapter