Classification of Intrusion Detection using PSO-SVM and Improved Decision Tree

Intrusion Detection is an efficient way of detecting the abnormal behavior of packets in the network, Although in data mining there are various effective decision tree based algorithms are implemented for the classification and detection of Intrusions in KDDCup99 Dataset. Here an efficient technique is implemented for the classification and detection of Intrusions in KDDCup99 Dataset using Feature Selection and Decision Tree based algorithms. The Proposed methodology works in two Stages Feature Selection using Particle Swarm Optimization with Optimization of PSO by Support Vector Machine and then Classification of Intrusion using Horizontal Partition Decision Tree. The Proposed Methodology implemented is more efficient in comparison with Decision Tree based algorithms.