Research on low-cost intrusion detection method for Internet of Things based on PCA dimensionality reduction
LIU Ziyi1, SONG Huazhu2
1.Faculty of Artificial Intelligence in Education， Central China Normal University， Wuhan 430079， China；2.School of Computer Science and Artificial Intelligence， Wuhan University of Technology， Wuhan 430070， China
Abstract：With the widespread popularity of smart homes and IoT devices， users have an increasing need to identify potential network security threats. Low-cost intrusion detection research has become a research hotspot in the field of IoT security. In view of this， this paper is committed to exploring a low-cost IoT intrusion detection method based on machine learning， that is， an IoT traffic intrusion detection model （PCA-RLR model） that integrates multiple regularization methods and dimensionality reduction based on principal component analysis （PCA）， aiming to significantly improve the effectiveness of network security protection. This paper improves the robustness of the model by optimizing and integrating multiple regularization methods， and uses PCA method to refine high-dimensional data features and reduce dimensions， thereby constructing a two-classifier model that can effectively identify normal traffic and abnormal attacks to provide security warning. Experimental results show that the logistic regression model integrating multiple regularization methods and PCA shows excellent performance in IoT intrusion detection tasks. Among them， L2 regularization enhances the stability and generalization ability of the model； PCA significantly reduces the feature space dimension， causing only a small performance loss under low computational complexity； the simulation experiment also verified the adaptive solver （Adaptive Solver） effectiveness on different data set characteristics. Experimental results show that the low-cost IoT intrusion detection model （PCA-RLR model） proposed in this article achieves high detection accuracy and low false alarm rate on the test set. This research provides a new low-cost method for network intrusion detection， which is expected to be widely used in actual smart home and Internet of Things device security protection.
刘子毅,宋华珠. 基于PCA-RLR模型的低成本物联网入侵检测方法研究[J]. 华中师范大学学报(自然科学版), 2025, 59(6): 831-842.
LIU Ziyi,SONG Huazhu. Research on low-cost intrusion detection method for Internet of Things based on PCA dimensionality reduction. journal1, 2025, 59(6): 831-842.
2025, Vol. 59 
