Machine Learning Based Intrusion Detection System

System administrators use a network intrusion detection system (NIDS) to identify network security breaches inside their own firm. Building a clever and robust NIDS for irregular and capricious attacks, however, raises various challenges. One of the key subjects in NIDS research in recent years has been the application of machine learning understanding of strategies. This approach provides a network intrusion detection tool that effectively identifies several kinds of network intrusions, including Dos, U2R, R2L, Probe, and Normal.It employs twin support vector machines and decision trees. The trees serve to construct the decision tree for network traffic data. Then, to maximize the separation of the top nodes of the decision tree, the bottom-up merging approach is applied, hence minimizing error buildup during generation. Embedding twin support vector machines inside the decision tree allows one to subsequently use the network intrusion detection model. This performance assessment is based on network intrusion detection analysis datasets—namely KDD-CUP99 and NSLKDD.