A Swarm Negative Selection Algorithm for Email Spam Detection

Abstract

The increased nature of email spam with the use of urge mailing tools prompt the need for detector generation to counter the menace of unsolocited email. Detector generation inspired by the human immune system implements particle swarm optimization (PSO) to generate detector in negative selection algorithm (NSA). Outlier detectors are unique features generated by local outlier factor (LOF). The local outlier factor is implemented as fitness function to determine the local best (Pbest) of each candidate detector. Velocity and position of particle swarm optimization is employed to support the movement and new particle position of each outlier detector. The particle swarm optimization (PSO) is implemented to improve detector generation in negative selection algorithm rather than the random generation of detectors. The model is called swarm negative selection algorithm (SNSA). The experimental result show that the proposed SNSA model performs better than the standard NSA.