In this paper, propose to use computational intelligence models to improve intrusion detection system, the computational intelligence algorithms are used as preprocessing steps for selecting most significant features from network data. Two computational intelligence algorithms, namely Ant Colony Optimization (ACO) and Particle Swarm Optimization (PSO) are implemented to generate subset of relevant features. The computational intelligence approaches have been applied to optimize the classification of algorithms. The most significant features obtained from computational intelligence is fed into the classification algorithm. Novelty of this presents research of use computational intelligence algorithms namely Ant Colony Optimization (ACO) and Particle Swarm Optimization (PSO) for handling dimensionality reduction. The dimensionality reduction is obstructed time processing of classification algorithms. Three classification algorithms namely K-Nearest Neighbors (KNN), Support Vector Machine (SVM) and Naïve Bayes (NB) are implemented for intrusion detection system. Benchmark datasets, namely, KDD cup and NSL-KDD datasets are used to demonstrate and validate the performance of the proposed model for intrusion detection. From the empirical results, it is observed that the classification algorithm has improved the intrusion detection system with using computational intelligence algorithms. A comparative result analysis between the proposed model and different existing models is presented. It is concluded that the proposed model has outperformed of conventional models.
INTRODUCTION
Data Security has seen huge advancement over the most recent couple of decades. The technologies have been developed so that the accessibility of electronic information handling frameworks/ information systems got inside the range of independent venture and home clients. These information systems got interconnected through an overall net
work by and large known as web. Furthermore, Enormous development of, broad utilization of web has changed the information prepared and directed organizations over the web in the most recent decade [1]. To ensure the computers and the cost of harms which is caused by such unapproved get to, a compelling and proficient intrusion detection system should be utilized to be protected the security of information systems. Presently a-days intrusion detection systems have developed to be an essential part of system security foundation [2]. The idea behind using data mining approaches is to allow system to detect the unknown attacks and also their variations. The aim target of this presents research work is to use the computational intelligence models to improve the classification algorithms for detecting intrusion. The paper is organized as follows in Section 1 presented an introduction.