An intelligent network intrusion detection system for anomaly analyzer using machine learning for software defined networks.

A novel Intrusion Detection System (IDS) architecture is explained in this paper by development of anomaly analyzersand NSL-KDD datasets approach. In present situation this Software Defined Network (SDN) has grown rapidly over vast areas. As it has been a new innovative method, it has brought out high accurate promising results which may lead to high demand for internet architecture in coming generations. Intrusion Detection System (IDS)is designed with a new architecture which includes an anomaly analyzer detection module. Our architecture is developed on trained NSL-KDD datasets as it is an effective dataset. By using a centralized SDN networking area has become flexible. In this process of advancement in networks it also developed risk of dangerous threats and more vulnerable environment has been created which resulted in various threats and damages such as system paralysis, online transactional frauds, network breakdown, Systems paralysis, and different kinds of robberies. Implementation of intelligent ML based algorithms in Software Defined Network (SDN) in a network intrusion detection system (NIDS) has improved considerably from past few years. In our research various kinds advanced and classical tree based ML techniques such as Random forest, Decision trees and our proposed XGBoostprocedures are selected for demonstrating the threat detection and accuracy standards. The benchmark dataset called NSL-KDD dataset is used for training and testing our methods, as it is an efficient dataset. In our work we used NSL-KDD as a multi-class classification task. It is for detectionof attack and identification of type of attack such as PROBE,DDoS,U2R and R2L etc. Observations showed that the hybrid approach in this paper gives accurate results when compared with other individual approaches. [ABSTRACT FROM AUTHOR]