Intrusion Detection System Using Improved Convolution Neural Network

Network intrusion detection technology plays an important role in maintaining network security, the main work is to continuously detect the current network status, through the detection of abnormal behavior in the network state, timely warning to alert network managers. The timeliness and accuracy of the intrusion detection system(IDS) is critical to the availability and reliability of the current network. In response to the problems of high false alarm rate, low detection efficiency and limited functions commonly found in IDS, this paper first investigates the application of deep learning techniques to the field of network intrusion detection. With the ability of deep learning algorithms to automatically extract features from intrusion data and avoid the work of manually screening features, an intrusion detection method based on improved convolution neural networks is then proposed. The method is improved by introducing Inception module for optimal intrusion feature extraction based on the traditional convolution neural network. The inception module employs a parallel convolution structure with different filters, using convolution kernels of different sizes on each convolution line for multiple layer-by-layer operations and The various features of network intrusions in the data set are identified and clustered by means of stacking.