Intrusion Detection-A Comparative Analysis Using Classification Algorithms

Intrusion Detection System is gaining more popularity nowadays as everyone is keen on protecting their networks. It is used to identify various authentic and malicious activities in the system and network. A lot of research activities are taking place to protect the network from outsiders as well as insiders. Various soft computing techniques like Data Mining, Artificial Intelligence are used in Intrusion Detection System for identifying malicious activities. In this paper we have done a survey on four supervised machine learning algorithms: Decision Tree (J48), K-Nearest Neighbor (KNN), Naive Bayes (NB) and Support Vector Machine (SVM). We have shown a comparative analysis of these algorithms based on Accuracy, True Positive Rate (TPR) and False Positive Rate (FPR). We have used NSL-KDD dataset for our experiment. Based on the experimental result, we have shown that the performance of Decision Tree (J48) and K-Nearest Neighbor are better than other two algorithms in terms of Accuracy, True Positive Rate (TPR) and False Positive Rat (FPR).