iSPmA: A Novel IOT Security Event Perception Model based on Autonomic Computing

According to the high dimension of security event feature and the difficulty of security event autonomic perception in the age of big data, a novel Inte rnet of Things ( IOT , for short) security event perception model based on Autonomic Computing and Principal Component Analysis (PCA, for short) is proposed, including element extraction, element understanding and event prediction. In which, to improve the real-time performance of element understanding, PCA is adopted to map the initial high dimensional feature to a set of new unrelated synthesized feature, and back propagation neural network (BP neural network, for short) is used to fuse the synthesized feature after reduction. The experimental result show that, feature reduction by PCA can greatly reduce the input dimension of fusion engine, efficiently cuts down the learning time of BP neural network, and improves the accuracy of event perception.