支持向量机的蛋白质远程同源检测方法分析  

Analysis of Protein Homology Remote Detection Methods Based on Support Vector Machine

作　　者：杨静[1] 殷志祥[1] 崔健中[1] 

机构地区：[1]安徽理工大学理学院,安徽淮南232001

出　　处：《安徽理工大学学报（自然科学版）》2009年第3期64-68,共5页Journal of Anhui University of Science and Technology:Natural Science

基　　金：国家自然科学基金资助项目(60274026;30570431;60873144);安徽省优秀青年基金资助项目(06042088);安徽省教育厅自然科学重点资助项目(2006kj068A);安徽省人才基金资助项目

摘　　要：支持向量机是目前蛋白质远程同源检测应用最成功的方法。在介绍这些基于支持向量机核方法的原理之后,比较这些检测方法的不同之处;再从复杂性角度对比分析不同方法的计算效率;最后指出核方法中核函数的选取也决定支持向量机的分类能力。Support Vector Machine is the most successful method of protein homology remote detection. After principle of these kernel methods based on support vector machine was presented, differences between these detection methods were compared. Based on viewpoint of complexity, comparative analysis of different calculation methods＇ efficiency was done. It＇s concluded that support vector machine classification ability depends on selected kernel function in kernel methods.

关 键 词：蛋白质远程同源检测 支持向量机 核函数 

分 类 号：Q-332[生物学]