Network Intrusion Detection with 1D Convolutional Neural Networks

Computer network assets expose to various cyber threats in today’s digital era. Network Anomaly Detection Systems (NADS) play a vital role protecting the purview of security. Intrusion detection systems data are imbalanced and high dimensioned, affecting models’ performance classifying malicious traffic. This paper uses denoising autoencoder (DAE) for feature selection reduce dimension. To balance data, authors use combined approach oversampling technique, adaptive synthetic (ADASYN) cluster-based under-sampling method using clustering algorithm, Kmeans. Then, one-dimensional convolutional neural (1D-CNN) is used perform classification. The proposed model evaluated on UNSW-NB15 NSL-KDD datasets. experimental results show that produces rate 98.79% 97.23% binary classification multiclass classification, respectively. In evaluation NSL-KDD, yields 98.52% type 98.16%