Intrusion Detection System using Hybrid Reverse - Binary Meerkat Optimization Algorithm and Support Vector Machine with Improved Random Forest

The rapid advancement of technology and the growth of the internet have made network security a major concern, with various attacks being addressed using Machine Learning (ML).However, as the demand for network systems continues to increase, existing approaches struggle to process relevant features, impacting performance.This research proposes a Hybrid Reverse Strategy with a Binary Meerkat Optimization Algorithm (HRMOA) combined with a Support Vector Machine and Improved Random Forest (SVMIRF) to efficiently analyze various attacks in an Intrusion Detection System (IDS) and achieve better accuracy.Data collected from the UNSW-NB15 and NSL-KDD datasets is pre-processed using Min-Max normalization to ensure that all features are treated equally, preventing bias caused by differences in feature scales.Feature selection using MOA efficiently finds the dynamic search space by utilizing the hybrid reverse strategy, avoiding local optima while ensuring the exploration phase identifies optimal features.The SVM-IRF method for classification effectively manages complex feature boundaries and highdimensional data, integrating with RF to ensure decision-making and achieve better classification accuracy.The proposed method achieves an accuracy of 99.36% on the UNSW-NB15 dataset and 99.24% on the NSL-KDD dataset when compared with existing Long Short-Term Memory (LSTM) technique.