Enhancing Network Security through Multiclass SVM-Based Intrusion Detection Systems

Abstract

This paper provides an intrusion detection system based on support vector machines (SVMs) for internet-based attacks on computer networks. The purpose of intrusion detection systems (IDS) is to anticipate and stop present and potential threats. SVMs are employed to detect and forecast anomalous system activity. The DARPA and KDD'99 intrusion detection evaluation datasets provided the training and testing data. Real-world data experimentation revealed encouraging outcomes for multiclass support vector machine intrusion detection systems.