A Deep Learning-Based Framework for Feature Extraction and Classification of Intrusion Detection in Networks

An intrusion detection system, often known as an IDS, is extremely important for preventing attacks on a network, violating network policies, and gaining unauthorized access to network. The effectiveness of IDS highly dependent data preprocessing techniques classification models used enhance accuracy reduce model training testing time. For the purpose anomaly identification, researchers have developed several machine learning deep learning-based algorithms; nonetheless, accurate with low test train times remains challenge. Using hybrid feature selection approach neural network- (DNN-) based classifier, authors this research suggest enhanced system (IDS). In order construct subset reduced optimal features that may be classification, consists three methods, namely, chi square, ANOVA, principal component analysis (PCA), applied. These methods are referred “the big three.” On NSL-KDD dataset, suggested receives then evaluated. proposed method was successful in achieving following results: reduction input by 40%, average 99.73%, precision score 99.75%, F1 99.72%, time 138% 2.7 seconds, respectively. findings experiments demonstrate superior performance other comparison approaches.