International Journal of Network Security & Its Applications

In intrusion detection systems, classifiers still suffer from several drawbacks such as data dimensionality and dominance, different network feature types, impact on the classification. this paper two significant enhancements are presented to solve these drawbacks. The first enhancement is an improved selection using sequential backward search information gain. This, in turn, extracts valuable features that enhance positively rate reduce false positive rate. second transferring nominal numeric ones by exploiting discrete random variable probability mass function problem of latter combined known normalization methods achieve a hybrid approach. Finally, intensive comparative study approves efficiency shows better performance comparing other proposed methods.