Network intrusion detection and classification using machine learning predictions fusion

The primary objective of an intrusion detection system (IDS) is to monitor the network performance and look into any indications malformation over network. While providing high-security IDS played a vital role for past couple years. will fail identify all types attacks, when it comes anomaly detection, often connected with high false alarm rate accuracy very average. Recently, utilize machine learning methods, because way that algorithms demonstrated have capacity adjusting as well permitting proper reaction real-time data. This work proposes prediction-level fusion model classification using techniques. also retraining unknown attacks increase effectiveness in IDS. experiments are carried out on security layer knowledge discovery database (NSL-KDD) dataset Konstanz information miner (KNIME) analytics platform. experimental results showed 90.03% simple 96.31% re-trained models. result inspires researchers use techniques build