An Intelligent Intrusion Detection System using Slime Mould Algorithm with Ensemble Machine Learning

Nowadays, real-world process has shifted to cyber environment specifically the computers are communicating with each other over the Internet. As result, it rising number of security exposures and network administrators are unable to secure network from all cyberattack forms. Therefore, this research proposes a Slime Mould Algorithm with Ensemble Machine Learning (SMA-EML) for Intrusion Detection System (IDS). The SMA helps to select relevant features while removing irrelevant one thereby reducing data dimensionality. SMA improves the feature set to demonstrate which attributes are used for enhancing IDS classification and differentiate between normal and abnormal traffic. The main goal of EML classifier is to apply a group of classifiers and ensure that it functions improved as a collective unit than as individual components. The EML classifier trains more classifiers and integrates the results by a voting system which enhanced generalization in classification of the attacks to facilitate mapping of the complex attack space. The hash encoding is used to convert categorical feature to an integer format that relies on hash functions and it maps the values. The proposed SMA-EML attains 99.75% f1-score, 99.85% accuracy, 99.82% recall and 99.79% precision for CICIDS-2018 dataset.