Improving accuracy for anomaly based IDS using signature based system

Anomaly based systems have become a vital information technology fields. In the evolution of anomaly based IDS, improving detection accuracy is more important. Anomaly based approach is efficient from signature based on computer network. But, in some case, signature based system is quickly identified attack from anomaly systems. In the last few years, massive data processing in computer network system. Therefore, reduce the data dimension in the stage of preprocessing, efficiency of detection can be improved much. We applied preprocessing in the KDD and our collected dataset using information gain. After that Naïve Bayes and Snort are used to classify the compression results and training the machine. This method can accomplish detection of network anomaly.