SYNERGISTIC INTRUSION DETECTION THROUGH INTEGRATION OF PARTICLE SWARM OPTIMIZATION ENHANCED WAVELET TRANSFORMS WITH OPTIMIZED LONG SHORT TERM VARIANTS

Abstract

This research presents an advanced intrusion detection framework utilizing a synergistic combination of preprocessing techniques and sophisticated machine learning models. The proposed methodology incorporates Particle Swarm Optimization (PSO), Discrete Wavelet Transform (DWT), and Stationary Wavelet Transform (SWT) as preprocessing algorithms to refine and prepare network datasets, ensuring optimal quality for subsequent analysis. Long Short-Term Memory (LSTM) networks, known for their effectiveness in sequence modelling, are employed for feature extraction and classification. Additionally, a proposed algorithm, Human Felicity Optimization LSTM (HFO-LSTM), is introduced, demonstrating a novel contribution to intrusion detection. The proposed HFO-LSTM algorithm is highlighted for its exceptional accuracy, marking a significant advancement in intrusion detection capabilities.