Comparison of Data Mining Techniques for Building Network Intrusion Detection Models

Harsha Kosta, Darshan Bhavesh
2016 International Journal of Computer Applications  
Intrusion detection is a detection of encroachment on the personal network or the private network to breach the security systems. This system provides analytical measures to gather information from various networks or computers to identify the cracks in the security systems caused by intruders. The sudden tremendous growth in the amount of internet users network intrusion detection has gained a huge amount of attention/need towards the research of network. Today, cyberattacks have become a
more » ... issue for any organization or individual in the network against preserving significant data and information in their personal computers connected to a network. In this paper, a comparative study was done on two different data mining techniques: decision tree and support vector machine algorithms. These techniques are implemented on the dataset for the experiment, since decision tree C5.0 technique and support vector machine (SVM) in general widely used in intrusion experiment data i.e. KDD CUP99 data set downloaded from UCI repository site. The better performance of C5.0 algorithm in terms of accuracy, sensitivity and specificity error measures are to be proved in this paper. General Terms
doi:10.5120/ijca2016909840 fatcat:skxh3gbxkvdp7dstnkqdbgovva