Rough Set Weighted Naïve Bayesian Classifier in Intrusion Prevention System

Effective action classify plays an important role in intrusion prevention system (IPS), especially application layers IPS. Classify is the basic in data mining tasks. Naive Bayesian classifier is widely used in data mining tasks due to its computational efficiency and competitive accuracy. For the naive Bayesian independent assumption is rarely realistic, and somehow depresses the classifierspsila performance. Many improved method was developed. This paper described a novel feature weighted naive Bayesian classifier using rough set upper approximation as a feature weighting coefficient. Experiment shows that RWNB has improved computing efficiency and higher interaction speed, and is a good choice for classifier in intrusion prevention system.