Web Log Intrusion Detection System by Using Fuzzy Approach A Review

The rapid growth of web-based applications has increased the exposure of web servers to diverse cyber threats. Web server logs provide valuable information for detecting malicious activities; however, traditional intrusion detection systems (IDS) often struggle with uncertainty, imprecision and evolving attack patterns. This paper presents an Extended Modified Fuzzy Possibilistic C-Means (EMFPCM) algorithm for web log intrusion detection, addressing limitations of existing fuzzy IDS methods in handling large datasets, overlapping attacks, noise and computational complexity. A comprehensive review of fuzzy-based IDS techniques, including fuzzy rule-based systems, fuzzy C-Means variants, Type-2 fuzzy systems, fuzzy neural networks, fuzzy support vector machines and hybrid fuzzy–deep learning approaches, was conducted. EMFPCM integrates fuzzy membership, possibilistic typicality and adaptive weighting to detect anomalies efficiently in large-scale web logs. Comparative analysis using metrics such as accuracy, computational complexity, scalability and noise resistance demonstrates that EMFPCM achieves high accuracy (92–95%), strong noise resistance, moderate complexity and high scalability, outperforming most traditional fuzzy IDS methods. The algorithm provides a balanced, practical solution, enabling real-time, robust anomaly detection while reducing false positives and computational overhead. EMFPCM demonstrates that combining fuzzy and possibilistic clustering with adaptive weighting offers an effective trade-off between accuracy, scalability and noise resilience, making it suitable for enterprise-level web security systems