A Novel Ensemble of Classification Techniques for Intrusion Detection System

Abstract

Massive volumes of data have been generated over time by various entities. This phenomenon is due to computer advancements. By recognizing attacks and employing acceptable practices, organizations have made the process of data inquiry extremely difficult. While Intrusion Detection Systems (IDSs) protect resources from threats, they are not up to the task of improving area precision, lowering false alarm rates, and identifying subtle threats. This paper proposes a design that combines data mining gathering estimations and association rules to lead network interference recognition. Tests are conducted to evaluate a few AI classifiers that rely on the KDD99 interference dataset. Among the information mining calculations, simple Bayes, decision trees, support vector machines, decision tables, k-nearest neighbor computations, and neural organizations are among the information mining calculations covered. The focus of the false negative and false positive execution metrics is to increase the detection speed of the intrusion detection network. This proposed work performs a better job in detecting intrusions than the previous strategies. This proposed organizational work is more effective than past methods at spotting anomalies.