A Hierarchical Intrusion Detection Model in Wireless Sensor Networks

Abstract

Aiming at the problems of poor detection performance and high model complexity of existing detection algorithms in wireless sensor networks (WSNs), a hierarchical intrusion detection model for wireless sensor networks is proposed. Firstly, the traffic data is preprocessed at ordinary nodes, and the chi-square test is used for feature selection to reduce the amount of data storage and calculation; secondly, the improved random forest classifier is deployed to the cluster head nodes; finally, the base station uses Light Gradient Boosting Machine to detect suspicious traffic data. Experimental results show that compared with the existing detection models, this model has lower model complexity and good detection performance.