Research on Intrusion Detection Algorithm Based on Deep Belief Networks and Three-way Decisions
Xiangtong Du, Yongzhong Li, Shipeng Zhang · Proceedings of the 2020 4th International Conference on Electronic Information Technology and Computer Engineering · 2020
Intrusion detection is one of the key technologies in network security. With the diversification and intelligence of network intrusion behaviors, it is difficult for traditional intrusion detection algorithms to abstract the features contained in the intrusion behaviors, and there are some shortcomings in the performance of intrusion detection. Therefore, this paper proposes an intrusion detection algorithm based on deep belief networks and three-way decisions. Firstly, using deep belief network to extract features from high-dimensional data, and a multi-granularity feature space is constructed after multiple feature extraction. Then, three-way decisions theory is used to classify the data in positive and negative domains according to the threshold, the classifier based on the three-way decisions theory makes immediate decisions on intrusion or normal data. For the uncertain network behaviors in the boundary domain, using KNN classifier according to the characteristics of different granularity to do further analysis. Experiments on KDD CUP 99 and NSL-KDD data sets show that this algorithm can improve the performance of intrusion detection.