An Approach to Feature Selection in Intrusion Detection Systems Using Machine Learning Algorithms

The rapid development of various services that are provided by information technology has been widely accepted by the users who are making use of such services in their day-to-day life activities. Securing such a system application from various intrusions still remains to be a one of the major issues in the current era. Detecting such anomalies from the regular events involves various steps such as data pre-processing, feature selection, and classification. Many of the computational models intend to accurately discriminate the samples of each group for better classification by identifying candidate features prior to the learning phase. This research studies the implementation of a combined feature selection technique such as the GRRF-FWSVM method which is applied to the benchmarked anomaly detection dataset KDD CUP 99. The results prove the novel proposed hybrid model is an effective method in identifying anomalies and it increases the detection rate of about 98.55% of the intrusion detection system with the two most common benchmark models.