Classification Method for Network Security Data Based on Multi-featured Extraction
Yunchuan Kang, Jing Zhong, Ruofeng Li, Yuqiao Liang, Nian Zhang
Abstract
A method of classifying network security data based on multi-featured extraction is proposed to address instability of a nonlinear time series in a network security threat. Cybersecurity information is divided in line with the principle of acquiring multiple attributes. On this basis, an adaptive adaptation estimation technology is optimized in analogue. With the proposed method, a cybersecurity information classification system is constructed according to the phase interval reconstruction principle so that a dynamic and autonomous adaptation estimation of the cybersecurity threat can be completed to ensure the feasibility of cybersecurity information classification. The experimental result proves that the cybersecurity information classification technology based on multi-attribute extraction can effectively guide chaos into adjacent orbits and reasonably control the training scale. Moreover, the accuracy of the estimation is guaranteed and the cybersecurity threat is estimated because of its high-speed convergence and strong proximity. Therefore, the proposed classification technology can assist professionals and backstage managers in guaranteeing security by facilitating receipt of information in a timely manner.