OLGBM: Optuna Optimized Light Gradient Boosting Machine for Intrusion Detection

Network technology has been evolved exponentially in the past few decades. At the same time, gazillions of network intrusion incidents are continuously forming cyberspace a shocking vulnerable domain to explore for the personnel from armature to the professionals. Consequently, networks mostly become botnet when there are no Intrusion Detection Systems (IDS) deployed. A smart and efficient intrusion detection system is an inexorable solution designed for fine-tuning the preventive rule-sets in any network. In this paper, we have proposed an efficient anomaly-based IDS mechanism. This detection mechanism has been competent by three tree-based state-of-the-art machine learning classifiers, namely, Random Forest (RF), Decision Tree (DT), and Optuna based Light Gradient Boosting Machine (OLGBM) algorithms. A comparative study of the algorithmic performances has been executed to determine the best algorithm that could be efficient for the IDS. In the experiment, the proposed OLGBM model gives better performance (accuracy 98.46%).