基于集成学习和多聚类的混合威胁检测算法

With the rapid development of information technology, the Internet of Things has driven explosive growth in traffic and has profoundly affected people\'s production and life through the interconnection of all things. At the same time, the information security issues brought about by the large-scale application of the Internet of Things have become increasingly prominent. However, the existing Internet of Things threat detection models generally have problems such as low detection accuracy, weak generalization ability, and difficulty in deployment, making it difficult to form an effective security defense. In order to address the above problems and challenges, this paper proposes a hybrid threat detection model based on ensemble learning and cluster combination model. Through comprehensive feature engineering, data dimension reduction and key feature screening are achieved, and a hierarchical detection framework is constructed with ensemble learning model and cluster combination model as the core, so as to achieve efficient detection of various known and unknown types of abnormal traffic. At the same time, an error corrector is constructed to reduce the overall false alarm rate of the model, and the Bayesian optimization method is used to achieve dynamic optimization of parameters of each layer model. Performance verification on three Internet of Things datasets shows that compared with the comparison model, the detection performance of the model is significantly improved, and it is more lightweight.