Date of Publication :24th January 2018

Abstract: In the present scenario of network architecture, handling long payload features is a challenge, specifically because many machine learning algorithms are not able to process these long payload features. Some of the Network Intrusion Detection Systems (NIDS) are completely avoiding these long payload features. To address this challenge, a new methodology called feature extraction through Bigram and Trigram techniques has been proposed. The long payload features are encoded though these proposed techniques and are prepared to be used in machine learning algorithms. Experiments were carried out on ISCX 2012 data set. The designed feature selection based system has shown a noticeable improvement on the performance using different metrics.

Reference :