Learning intrusion detection based on adaptive bayesian algorithm

Recent intrusion detection have emerged an important technique for information security systems. Its very important that the security mechanisms for an information system should be designed to prevent unauthorized access of system resources and data. Last few years, many intelligent learning techniques of machine learning applied to the large volumes of complex and dynamic audit data for the construction of efficient intrusion detection systems (IDS). This paper presents, theoretical overview of intrusion detection and a new approach for intrusion detection based on adaptive Bayesian algorithm. This algorithm correctly classify different types of attack of KDD99 benchmark intrusion detection dataset with high detection accuracy in short response time. The experimental result also shows that, this algorithm maximize the detection rate (DR) and minimized the false positive rate (FPR) for intrusion detection.