Clustered negative selection algorithm and fruit fly optimization for email spam detection

摘要

At present, spam is an actual and increasing problem that compromises email communications across the world. Thus, several solutions have been proposed to stop or reduce the amount of this threat. However, methods based on negative selection algorithm (NSA) lack continuous adaptability and suffer from low detection performance. Moreover, these methods require a large number of detectors to cover all non-self spaces. Thus, this study proposes a new e-mail detection approach based on an improved NSA called combined clustered NSA and fruit fly optimization (CNSA-FFO). The system combines actual NSA with k-means clustering and FFO to enhance the efficiency of classic NSA. Experiments results in spam benchmark show that the performance of CNSA-FFO is better than the classic NSA and NSA-PSO, especially in terms of detection accuracy, positive prediction, and computational complexity.