Research Article
PSVM: Quantitative Analysis Method of Intelligent System Risk in Independent Host Environment
@INPROCEEDINGS{10.1007/978-3-030-48513-9_43, author={Shanming Wei and Haiyuan Shen and Qianmu Li and Mahardhika Pratama and Meng Shunmei and Huaqiu Long and Yi Xia}, title={PSVM: Quantitative Analysis Method of Intelligent System Risk in Independent Host Environment}, proceedings={Cloud Computing, Smart Grid and Innovative Frontiers in Telecommunications. 9th EAI International Conference, CloudComp 2019, and 4th EAI International Conference, SmartGIFT 2019, Beijing, China, December 4-5, 2019, and December 21-22, 2019}, proceedings_a={CLOUDCOMP}, year={2020}, month={6}, keywords={DT-SVM PSO-SVM Ubiquitous network}, doi={10.1007/978-3-030-48513-9_43} }- Shanming Wei
Haiyuan Shen
Qianmu Li
Mahardhika Pratama
Meng Shunmei
Huaqiu Long
Yi Xia
Year: 2020
PSVM: Quantitative Analysis Method of Intelligent System Risk in Independent Host Environment
CLOUDCOMP
Springer
DOI: 10.1007/978-3-030-48513-9_43
Abstract
Quantitative risk analysis of security incidents is a typical non-linear classification problem under limited samples. Having advantages of strong generalization ability and fast learning speed, the Support Vector Machine (SVM) is able to solve classification problems in limited samples. To solve the problem of multi-classification, Decision Tree Support Vector Machine (DT-SVM) algorithm is used to construct multi-classifier to reduce the number of classifiers and eliminate non-partitionable regions. Particle Swarm Optimization (PSO) algorithm is introduced to cluster training samples to improve the classification accuracy of the constructed multi-classifier. In the ubiquitous network, the cost of information extraction and processing is significantly lower than that of traditional networks. This paper presents a quantitative analysis method of security risk based on Particle Swarm Optimization Support Vector Machine (PSO-SVM), and classifies the flow data by combining the way of obtaining the flow data in ubiquitous networks, so as to realize the quantitative analysis of the security risk in ubiquitous networks.
