Feature Reduction and Classifications Techniques for Intrusion Detection System
Gulab Sah, Subhasish Banerjee
Abstract
Today, an intelligent intrusion detection system is very important to enable high-level security in Networks to protect private and highly sensitive information. Nowadays, an intrusion detection system gaining more interest to the researcher, because of the rapid increase in the use of internet and network technology which also led to growth in the number of attacks. These attacks should be prevented and detected by the intelligent intrusion detection system. To deal with these attacks the intelligent intrusion detection system uses large datasets; however have a curse of dimensionality. To handle these high dimensionality dataset intrusion detection systems uses feature reduction methods for achieving less time complexity and reduces resource utilization. In this paper, we have discussed the various types of intrusion detection system, reduction methods and classification techniques of machine learning for intelligent intrusion detection system such as K-nearest neighbors, Support vector machine, Random Forest, and Naive Bayes. The main purpose of this paper is to propose a method that will determine, whether or not with the selected features, the accuracy rate will be improved or not compare to the accuracy rate with all features.