Optimized Spam Classification Approach with Negative Selection Algorithm

Abstract

This paper initializes a two element concentration vector as a feature vector for classification and spam detection. Negative selection algorithm proposed by the immune system in solving problems in spam detection is used to distinguish spam from non-spam (self from non-self). Self concentration and non-self concentration are generated to form two element concentration vectors. In this approach to e-mail classification, the e-mail are considered as an optimization problem using genetic algorithm to minimize the cost function that was generated and then classification of these cost function shall aid in creating a classifier. This classifier will aid in the new formation of algorithm that comprises of both greater efficiency detector rate and also speedy detection of spam e-mail. The algorithm implementation of the research work shall come in stages were spam and non-spam are detected in all phases for an efficient classifier.