New Hybrid Intrusion Detection System based on Data Mining Technique to Enhanced Performance

Intrusion Detection Systems (IDSs) is an efficient defense technique against network attacks as well host attacks since they allow network/host administrator to detect any type policy violations. However, Standard IDS are vulnerable and they are not reliable to novel and original malicious attacks. Also, it is very inefficient to analyze from a big amount of data such as possibility logs. Moreover, there are high false positives and false negatives for the common OSs. There are many other techniques which can help to improve the quality and results of IDS in which data mining one of them where it has been popularly recognized/identify as an important way to mine useful information from big amount of data which is noisy, and random. Integration of various data mining techniques with IDS to improve efficiency is the motive of proposed research. Proposed research is combining three data mining technique to reduce over head and improve execution efficiency in intrusion detection system (IDS). The Proposed research that ensembles clustering (Hierarchical) and two classifications (C5.0, CHAID) approaches. Proposed IDS execute on the standard KDD’99 (knowledge Discovery and Data Mining) Data set; this data set is used for measuring the performance of intrusion detection systems. Proposed system can detect the intrusions and classify them into four categories: Probe, Denial of Service (DoS), U2R (User to Root), and R2L (Remote to Local). A presented experiment results is carried out to the performance of the proposed IDS using KDD 99’ dataset. Its shows that the proposed IDS performed better in term of accuracy, and efficiency.