ChOsLSTM: Chebyshev Osprey Optimization-Based Model for Detecting Attacks

This paper explores the application of Chebyshev Osprey optimization-based Long Short Term Memory (ChOsLSTM) for intrusion detection, utilizing input from the CICIDS-2018 dataset. Initially, the dataset undergoes balancing using SMOTE to mitigate biased outcomes, followed by Z-Score normalization for data standardization. Subsequently, the ChOsLSTM model is employed, wherein the adjustable parameters of LSTM are fine-tuned using the ChOs algorithm. Notably, the ChOs algorithm incorporates Chaotic Chebyshev mapping within the conventional Osprey algorithm, enhancing randomness to prevent the algorithm from getting trapped at local optimal solutions.