Intrusion Detection Based on IDBM

With the advancement of internet, computer network security has brought serious concerns. Intrusion detection is an important topic in network security framework. To address the effectiveness and efficiency problem with traditional intrusion detection models, we present an intrusion detection method based on deep leaning. The deep belief network (DBN) constructed via the stacked Boltzmann machine model (RBM) is selected. Firstly, combining numeralization of symbolic features and numeric features normalization are used to processing network log features. In addition, extreme learning machine (ELM) was applied into the learning process of DBN. Compared with traditional DBN, the experimental results on the NSL KDD dataset demonstrate that intrusion detection based on IDBN has double training speed compared to DBN, while achieving a reliable detection rate.