Abstract: With the tremendous growth of the usage of computers over network and development in application running on various platform captures the attention toward network security[1]. Intrusion detection system has become an important component of a network infrastructure protection mechanism. The Intrusion Detection System (IDS) plays a vital role in detecting anomalies and attacks in the network [5]. In this work, data mining concept is integrated with an IDS to identify the relevant, hidden data of interest for the user effectively and with less execution time. In proposed system, we first preprocess dataset (KDD 99 cup).  Then we study different types of decision tree algorithms (C4.5 and its extension) of data mining for the task of detecting intrusions and compare their relative performances. Based on this study, it can be concluded that even extended C4.5 is complex but decision tree obtained is the most suitable with high true positive (correct detection of attacks) and low false positive (Incorrect detection) with high accuracy.

 

Keywords: Intrusion detection system, KDD 99 cup, Data Mining, Decision Tree Algorithms, C4.5 and its extensions