A New Intrusion Detection System Based on Rough Set Theory and Fuzzy Support Vector Machine

Nowdays, IDS (Intrusion Detection System) is a hot topic in the information security. The main function of IDS is distinguishing and predicting normal or abnormal behaviors. This paper is to propose a model used on IDS, it is based on rough set (RS) theory and fuzzy support vector machine (FSVM). Firstly, the model set rough set as a preprocessor of FSVM. Rough set can reduce dimensions of attributes and filter some invasion behaviors which are esay to identify. Secondly, less attributes selected by RS are input FSVM to train and classify, this method can improve operational speed of FSVM. For this model, FSVM uses an effective Fuzzy Membership Function based on the affinity among sample points to select an appropriate fuzzy membership to reduce the effects of outliers. Finally, Experimental results will show that the RS-FSVM performs the best recognition ability, indicating that RS-FSVM can serve as a promising model for intrusion detection system.