Escalation of the internet and the demand for magnifying the network security is increasing exponentially. The intrusion detection is an active network monitoring practice for finding the unauthorized access, policy violations, anomalous behaviors, malicious attacks, unconcerned packets detection etc. The present security fraternities like antivirus, cryptography and firewalls could not ensure complete protection for the systems linked with the internet. This paper reviews various data mining based intrusion detection techniques. Deep emphasis is given to observe the best among the Machine learning algorithm that records the optimal degree of attack detection and false alarm rate. The analysis is made using KDD cup 99 dataset. The algorithms like J48, Random tree, and Random forest are evaluated and equated to identify their detection ratio.

Keywords: IDS-Intrusion Detection System; MBID ABID,HBID, NBID -(Misuse, Anomaly, Host, Network) Based Intrusion Detection; Packet Sniffing; Feature selection; Outlier detection; False Alarm; J48; RF-Random Forest; RT-Random Tree.;