高维数据中的相似性度量算法的改进  

Optimization of Algorithm of Similarity Measurement in High-Dimensional Data

作　　者：邵昌昇[1] 楼巍[1] 严利民[1] 

机构地区：[1]上海大学机电工程与自动化学院,上海200072

出　　处：《计算机技术与发展》2011年第2期1-4,共4页Computer Technology and Development

基　　金：上海市(科委)"科技创新行动计划"非政府间国际科技合作项目(09530708600)

摘　　要：高维数据之间的相似性度量问题是高维空间数据挖掘中所面临的问题之一。为了有效解决高维效应给相似性度量带来的种种问题,首先分析传统相似性度量算法,得出其局限性。再通过对传统度量算法进行改进,提出新的Close函数,以弥补传统相似性度量算法应用在高维空间时的不足。提出Close函数后,将其与几种传统的相似性度量算法作比较,得出新算法在高维空间相似性度量方面的优越性。文中最后用Matlab对该函数做了定量分析,实验证明该函数在高维空间中能有效避免噪声和维灾效应的影响。The problem of similarity measurement between high dimensional data is one of the problems high-dimensional data mining faces.In order to solve the problems of high-dimensional similarity measurement,analysis of traditional algorithms are made at first to obtain limitation.A new function Close（） is presented based on the improvement of traditional algorithm to make up for the inadequate of traditional algorithm used in high-dimensional space.Advantages of the new function are obvious in high-dimensional similarity measurement after the comparison between Close（） and tradition algorithms are made.Quantitative analysis of function Close（） is made with Matlab and experiments prove that this function can avoid the affects of noise and the curse of high-dimension.

关 键 词：数据挖掘 高维数据 相似性度量 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]