Intrusion Detection System Using Machine Learning Techniques

Abstract

Network intrusion detection technology plays a vital role in guaranteeing network security. The primary goal is to continually monitor the current state of the network, identify abnormal behaviour in the network state, and notify network administrators in time. The speed and accuracy of an intrusion detection system (IDS) are important to the availability and dependability of today's network. In response to the difficulties of high false alarm rates, poor detection efficiency, and restricted functionality prevalent in IDS, this study first studies the application of machine learning approaches to network intrusion detection. Since machine learning algorithms can automatically extract characteristics from intrusion data and prevent human feature extraction, an intrusion detection approach based on a decision tree classifier is presented. The approach has been enhanced by integrating the Inception module for optimum separation of intrusion functions. The initialization module employs a classification structure with distinct filters, utilizing varying size classification kernels in each row to operate in several layers, and the diverse features of network incursions in the dataset are identified and categorized by stacking.