基于扩展范式距离的纵向数据相似性度量  

Algorithm based on norm distance distance similarity measurement for longitudinal data

作　　者：李会民[1] 方丽英[1] 闫健卓[1] 王普[1] 赵辉[1] 

机构地区：[1]北京工业大学,电子信息与控制工程学院,北京100124

出　　处：《计算机与应用化学》2012年第10期1176-1180,共5页Computers and Applied Chemistry

基　　金：北京杰出人才基金项目(2010D005015000001)

摘　　要：为解决纵向数据相似性比较问题,对基于扩展范式距离的纵向数据相似性度量方法进行了研究。使用基于粗糙集理论的核约简对属性变量进行选择,移去数据集中的冗余属性;用扩展范式距离进行数据项问的度量。为计算两数据项之间的相似性,把相关度特征值当作权重,通过扩展范数距离比较项与项相应主元之间的相似性。与其它3种度最方法的对比实验显示,所提出的纵向数据相似度测量方法是有效可行的,且在信息检索时的Recall与Precision优于其它同类方法。To solve the problem of the similarity measure of longitudinal data, this paper does research on similarity measure based on Extended Frobenius Norm （Eros） distance for longitudinal data. Feature selection via core reducing of rough set theory is used to delete the redundant attributes, and then we compare the similarity of item-pairs via Eros distance. For calculating the similarity degree of two items, paper put the eigenvalue as the weights; calculate the similarity between principal items by using Eros distance. By comparing experimental, conclusion can be drawn that this method can measure the distance of longitudinal data effectively and efficiently, and have good behaviors on recall and precision in information searching.

关 键 词：纵向数据 扩展范式距离 粗糙集 核约简 相似性度量 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]