Network intrusion detection using Deep Learning techniques
In this paper we compare the use of various machine learning techniques, to the use of various deep learning techniques for detecting intrusions in networks. The dataset that we used is NSL-KDD dataset, this dataset is a refined version of KDD’99 dataset. We utilize various supervised machine learning techniques and deep learning techniques to train and build multiple classification models that can classify attack type of network traffic versus normal type of network traffic. We compare the test accuracies of these multiple classification models to identify the best model for performing network intrusion detection. We conclude that using a combination of AutoEncoder for feature selection, followed by fully connected deep learning model for classification gave better accuracy compared to using machine learning models.