Dynamic Threshold-Based Algorithm for Client-Based HTTP Proxy Attack Detection through Spatial and Temporal Behavior Pattern Analysis

This paper provides a unique approach to client- based HTTP proxy attack detection using a Dynamic Spatiotem- poral Behavior Analysis (DSTBA) algorithm. Traditional meth- ods often lack adaptability to sophisticated cyberattacks. DSTBA addresses this by dynamically adjusting detection thresholds based on real-time analysis of spatial (network node distribution and interaction) and temporal (request timing and frequency) behavior patterns. This integration with machine learning tech- niques enhances attack identification accuracy while minimizing false positives. DSTBA’s core strength lies in detecting subtle deviations in client behavior indicative of proxy-based attacks. A feedback mechanism continuously refines thresholds, enhancing resilience against evolving attack strategies. Evaluations using real-world datasets exhibit DSTBA’s superior performance com- pared to static threshold methods, achieving high detection rates with low false alarms. This research contributes significantly to network security advancements by providing an adaptable and efficient solution for client-based HTTP proxy attack detection in the face of ever-changing cyber threats.