Streamlined Intrusion Detection: Faster, Accurate, and Complexity-Reduced System
Pratyashi Satapathy, Divya Sharma, Sampangi Rama Reddy B. R., Balaji K A, K. S. Bhuvaneshwari, Sahil Khurana · 2025
The complexity-reduced intrusion detection system is presented, which makes an effort to properly identify assaults that are taking place on a network with the assistance of the Back Propagation Neural Networks and the Aho-Corasick matching of patterns method. A Fuzzy Firefly procedure is used in this study to get rid of duplicated record sets, and an updated KNN-based imputed approach is provided with the assistance of the bagged methodology in order to deal with value gaps. After the dataset has been prepared, a genetic-based feature elimination process is carried out in order to exclude the characteristics that are not related to the input dataset. After the feature reduction has been completed, the optimum selection of characteristics is carried out with the help of the Hybrid Mosquito Swarm technique, and the Cuckoo Search method is applied to pick those characteristics from the input dataset that are the most optimal. In the final but not least, a BPNN and Aho-Corasick pattern recognition algorithms are used in order to perform Intruder Diagnosis. Outperforming approaches such as Tree-CNN, PM-RNN-IDS, and CNN-OR-IDS, the suggested approach Backpropagation Neural Network with Adaptive Clustering and Pattern Recognition Intrusion Detection System (BPNN-ACP-CR-IDS) obtains an overall accuracy of 95.2%. This proves that the suggested strategy is more effective and resilient when dealing with different percentages of attacker nodes.