An Approach to Improving Intrusion Detection System Performance Against Low Frequent Attacks

摘要

Network security is crucial in contemporary company. Hackers and invaders have regularly disrupted huge company networks and online services. Intrusion Detection Systems (IDS) monitor and report on harmful computer or network activities. Intrusion detection aims to detect, prevent, and react to computer intrusions. Researchers have suggested the fuzzy clustering-artificial neural network to improve intrusion detection systems. A hybrid Artificial Neural Network technique combines fuzzy clustering and neural networks to increase intrusion detection systems' accuracy, precision, and resilience. We built fuzzy clustering modified artificial neural networks to increase low-frequency attack detection and training time. This approach can be improved in terms of training duration and low-frequency attack accuracy. Our novel technique, Fuzzy Clustering-Artificial Neural Network-modified, beats the fuzzy clustering-artificial neural network algorithm by 39.4 in identifying low-frequent assaults and decreases the projected training time by 99.7.