The XGBoost Tuning by Improved Firefly Algorithm for Network Intrusion Detection

Abstract

Research proposed in this article presents a novel improved version of widely adopted firefly algorithm and its application for tuning the eXtreme Gradient Boosting (XGboost) hyper-parameters for network intrusion detection. One of the greatest issues from the domain of network intrusion detection systems are relatively high false positives and false negatives rates. In the proposed study, by using XGboost optimized with enhanced firefly algorithm, this challenge is addressed. Devised method was adopted and tested against recent benchmarking USNW-NB15 dataset for network intrusion detection. Achieved results of proposed method were compared to the ones obtained by standard machine learning methods, as well as to XGBoost models tuned by other swarm algorithms. Reported comparative analysis results prove that the proposed metaheuristics has significant potential in tackling machine learning hyper-parameters optimization challenge and that it can be used for improving classification accuracy, precision, recall, f1-score and area under the receiver operating characteristic curve for network intrusion detection datasets.