大规模曲线的自动分类方法及其应用  

Automatic Classification Method of Large-scale Curves and Application

作　　者：叶明[1,2] 王惠文[3] 寇薇[3] 

机构地区：[1]复旦大学公共管理博士后流动站,上海200433 [2]上海WTO事物咨询中心,上海200336 [3]北京航空航天大学经济管理学院,北京100083

出　　处：《系统管理学报》2010年第6期640-644,共5页Journal of Systems & Management

基　　金：国家自然科学基金创新研究群体科学基金资助项目(70521001);国家自然科学基金资助项目(70771004;70371007)

摘　　要：提出了一种针对大规模曲线的自动分类方法。该方法通过引入曲线类的相似精度作为算法的终止条件,根据相似精度的阈值将曲线集合自动地划分成若干满足精度要求的曲线类别。与传统的曲线聚类分析方法相比,该方法最重要的特点是可以自动确定分类的个数,而无须依赖分析人员主观猜测分类的个数或进行大量的试算。使用这一方法,可以在满足任意事先确定的精度要求下,自动得到曲线聚类的分析结果,因而大大提高了分析效率。An automatic classification method for large-scale curves is proposed.The method introduce similarity precision of curve as its termination condition,According to the threshold of similarity precision,the massive curves are decomposed into several classifications,which fulfill precision requirements.Compared with traditional methods for curve cluster analysis,this method can automatically determine the number of classifications,rather than decided by analysts subjectively.According to satisfied with the request of the precision which is predetermined,the application of such method enables the analysis result to be automatically generated,which in fact would substantially enhance efficiency.

关 键 词：曲线分类 大规模曲线 自动分类 

分 类 号：O212[理学—概率论与数理统计]