Data Mining Based Advanced Algorithm for Intrusion Detections in Communication Networks

Nowadays the network security is the important topic to research. The Network Security from different types of attacks which is R2L, U2R, and DoS. It is very challenging tasks due to variety of research problems like noise, large data size, inefficient features selection method etc. Network Intrusion Detection System (IDS), as the basic security protection technique, is generally used limiting such malicious attacks. In this project, we are presenting the efficient IDS solution using filter based feature choice strategy. We are exhibiting the hybrid feature determination method. The Intrusion Detection System (IDS) examine the main part in distortion and ambushes in the framework. In this examination work, data mining methods unite with association rule features extraction and classifier. In this paper, we proposed filter based hybrid feature selection algorithm (HFSA), most relevant features are retained and used to construct classifiers for respective classes. In this system it is worked on the real time packets, which is captured using the Jpcap library. Along with HFSA method, first we are contributing approach of detecting the cyber-attack brute force by modifying the algorithms of HFSA and classification. For the classification we used the Naïve bayes classifier. The performance of proposed method shows the efficiency compared to other methods.