文章摘要
张宇,刘雨东,计钊.向量相似度测度方法[J].声学技术,2009,(4):532~536
向量相似度测度方法
Vector similarity measurement method
投稿时间:2008-07-09  修订日期:2008-10-08
DOI:
中文关键词: 相似度  距离测度  相似性函数
英文关键词: similarity  distance measure  comparability function
基金项目:
作者单位E-mail
张宇 中国船舶重工集团公司第七六○研究所, 大连, 116013 zhangyuzy@163.com 
刘雨东 中国船舶重工集团公司第七六○研究所, 大连, 116013  
计钊 中国船舶重工集团公司第七六○研究所, 大连, 116013  
摘要点击次数: 1184
全文下载次数: 3955
中文摘要:
      判断两个向量的相似程度,一般采用计算相似度的方法。首先对相似度测度法中的几种成熟的计算方法和应用优缺点进行了综合分析和对比。然后,根据实际的应用需要,结合两种成熟的距离测度法的方法,提出了一种改进的距离测度,即相对误差距离;又从放大较大相对误差作用的思路出发,提出一个相似性函数,即最大相异系数。这两种方法都充分考虑了相对变化量,具有针对性,适用于未知的待比较向量与确知的标准向量进行测度的情况。
英文摘要:
      To estimate two vectors’ similarity, the methods of computing the similarity are generally adopted. Several mature similarity estimate methods and their merits and drawbacks in applications are analyzed and compared in this paper. According to the demand of practice application and associating with two mature distance measure methods, an improved distance measure method called relative error distance is proposed. And, from the thought of magnifying the effect of the larger relative error, a comparability function called most dissimilarity coefficient is created. These two methods, which have fully considered the relative variation value, are applicable in the case of measuring an unknown comparative vector and a known standard vector.
查看全文   查看/发表评论  下载PDF阅读器
关闭
function PdfOpen(url){ var win="toolbar=no,location=no,directories=no,status=yes,menubar=yes,scrollbars=yes,resizable=yes"; window.open(url,"",win); } function openWin(url,w,h){ var win="toolbar=no,location=no,directories=no,status=no,menubar=no,scrollbars=yes,resizable=no,width=" + w + ",height=" + h; controlWindow=window.open(url,"",win); } &et=E9D0CCCB93A79728D9E3423D16E4E5BFE2EA448C1DEC8E19E0DE292A42B61556F6E7FB9A3D857253&pcid=5B3AB970F71A803DEACDC0559115BFCF0A068CD97DD29835&cid=84529CA2B2E519AC&jid=DDCFCD5ACE1B1E5A6D46213553C850CA&yid=DE12191FBD62783C&aid=&vid=&iid=E158A972A605785F&sid=640CCB6E396307A8&eid=323E8A365B085E0B&fileno=20090421&flag=1&is_more=0">