文章提出基于二阶段法的凸壳支持向量机的研究,并探索其在网络服务器性能分析中的应用。实验表明,引入凸壳理论提升了支持向量机的执行效率,实现了对网络服务器性能参数的快速精确划分,对进一步研究基于模式识别的服务器性能检测与报警具有指导意义。 This paper presents the research of convex hull support vector machines based on two-stage method and explores its application in the performance analysis of network server. Experiments show that the introduction of convex hull theory enhances the execution efficiency of SVM and realizes the fast and accurate classification of network server performance parameters. It is instructive to further study the server performance detection and alarm based on pattern recognition.