一种文本数据挖掘与可视化的新方法  

A novel method of literature data mining and visualization

作　　者：宋驰[1] 刘青焱[1] 张颜[1] 蒋博[1] 徐振愚[2] 夏永静[1] 刘湘军[1] 

机构地区：[1]清华大学,北京100084 [2]北京林业大学,北京100083

出　　处：《北京生物医学工程》2008年第2期156-160,共5页Beijing Biomedical Engineering

基　　金：教育部重点项目(104232);教育部跨世纪人才基金;国家自然科学基金(90412018);清华裕元医学科学研究基金资助

摘　　要：传统搜索引擎的搜索结果采用的是以一维列表的形式展现,随着文献数据的急剧增多,用户对于搜索结果的辨识和分析速度也在急剧下降。为了弥补传统工具的这一不足,本文开发了"语义图"(semantic map,SMAP),此工具对文献数据进行数据挖掘和可视化,包括关联匹配和聚类,将搜索结果以二维矩阵图的形式展示出来,方便用户理解数据之间的内部联系,并帮助用户迅速从整体上把握搜索结果。最后以蛋白质组学文献分析过程为例具体展示了此工具的应用。The results of traditional search engine technologies are currently shown as one-dimensional list. With the explosion of publication data, users get rapid slowdown on the cognition of the results. Hence we were motivated to develop a novel method of literature data mining and visualization called semantic map（SMAP） , utilizing data mining tech- niques and visualization, including association matching and clustering, etc. A semantic map is shown as a two-dimensional array image to reveal the inner relations of data, and could help users quickly grasp the entire results from whole data. Finally, literature data of proteomics was analyzed, which was a good example of applying SMAP.

关 键 词：文献数据挖掘 可视化 语义图 

分 类 号：R318.04[医药卫生—生物医学工程]