Novel Intrusion Detection Algorithm of Information System based on RBM-KSVM

To improve the performance of traditional models, this paper proposes the novel intrusion detection algorithm of the information system based on RBM-KSVM which belongs to the machine learning based method. As network security has received attention from various institutions, research related to intrusion detection has also received attention from the academic community, the existing methods always suffer from the low accuracy. In this designed model, the information system data features are firstly studied. For this study, the representation is selected as virtual, while the characteristics are selected as the mixed data neighborhood space. Secondly, RBM can convert input to output and then reconstruct the input from the output as an unsupervised generative model that reconstructs original data. This will optimize the parameter of the detection model. In the experiment, the performance regarding kernel and accuracy are both tested. Through the results, it is evident that the proposed model is efficient.