Network Intrusion Detection system based on Feature Selection and Triangle area Support Vector Machine
Venkata Suneetha Takkellapati · 2012
As the cost of the data processing and Internet accessibility increases, more and more organizations are be- coming vulnerable to a wide range of cyber threats. Most current offline intrusion detection systems are focused on unsupervised and supervised machine learning approaches. Existing model has high error rate during the attack classification using support vector machine learning algorithm. Besides, with the study of existing work, feature selection techniques are also essential to improve high efficiency and effectiveness. Performance of different types of attacks detection should also be improved and evaluated using the proposed approach. In this proposed system, Information Gain (IG) and Triangle Area based KNN are used for selecting more discriminative features by combining Greedy k-means clustering algorithm and SVM classifier to detect Network attacks. This system achieves high accuracy detection rate and less error rate of KDD CUP 1999 training data set.