Intrusion detection system based on QBSO-FS

In traditional internet systems, intrusion detection is an effective way to ensure network security. However, due to the problem of computing resources, complex intrusion detection models are not suitable for the massive heterogeneous network environment of the Internet of Things. Appropriate data sets and appropriate classification algorithms are the keys of intrusion detection systems in the Internet of Things era. This paper proposes an intrusion detection system based on QBSO-FS and machine learning. The system is used to screen network feature data in actual industrial environments, and can establish an intrusion detection model. Experiments show that using a subset of the original data after feature selection can effectively improve the classification accuracy of ordinary machine learning algorithms. After fusing the optimal feature subsets classified by multiple machine learning algorithms, the data set performs better than a subset of general feature selection algorithms on high-performance machine learning classification algorithms.