DDoS Attack Detection Using Optimized Long Short-Term Memory Based on Improved Bacterial Foraging Optimization

Recent years have seen a notable rise in the variety of network attacks due to the widespread usage and expansion of the cyber world. One of the most significant threats to the Internet nowadays is a distributed denial-of-service (DDoS) attack, in which the attacker bombards the target system with a high number of packets in an attempt to prohibit regular users from accessing it. To protect against these threats, accurate attack detection measurement is essential. To identify abnormalities, an optimized long-short-term memory (LSTM) algorithm is proposed in this research. In optimized LSTM, the hyperparameters are optimized by using improved bacterial colony optimization (IBFO). IBFO uses the centroid opposition-based learning (COBL) method for initializing the population and position updating. The experimental result demonstrated that the modified LSTM method achieved excellent detection accuracy and quick convergence when compared to a few popular detection techniques.