基于同义词关系的个人文件搜索方法  

Searching Personal Files Based on Synonym Relationship

作　　者：赵喜燕[1] 李玉坤[1,2,3] 刘经雨 肖迎元[1,2,3] 冯美玲[1] 

机构地区：[1]天津理工大学计算机与通信工程学院,天津300384 [2]智能计算及软件新技术天津市重点实验室,天津300384 [3]计算机视觉与系统教育部重点实验室,天津300384

出　　处：《小型微型计算机系统》2014年第9期1971-1976,共6页Journal of Chinese Computer Systems

基　　金：国家自然科学基金面上项目(61170027;61170174)资助;国家自然科学基金青年基金项目(61202169)资助

摘　　要：用户在计算机中查找个人文件时,有时会因为记不住确切的文件名和存放位置而遇到困难,现有的基于关键词的桌面搜索工具不能帮助人们有效地解决这一问题.针对此问题,本文首先基于从多个用户搜集的个人文件集合,分析了个人文件名中所包含的词语之间的同义词关系,提出了个人文件同义词概念模型,提出并阐述了个人文件、个人词典、词语偏好度、同义词图、文件关键词向量等概念.基于该模型进一步提出了个人同义词词典与同义词图的增量构建方法,以及基于同义词的个人文件搜索方法.通过实验证明,本文提出的个人文件搜索方法在查全率、查准率和性能方面均优于传统的桌面文件搜索工具,同时在实验中也发现了个人文件管理行为中的一些规律.People sometimes cannot remember the exact words included in a file＇s name or its location and meet trouble in re-finding the file, when the existing keyword-based desktop search tools cannot work efficiently. Aiming at this problem, this paper first analyzes the synonymous relationship among the words included in desktop filenames based on a data set we collected from multiple users, and then puts forward a synonym conceptual model towards personal files. Moreover, it proposes and expounds the concepts of Personal File, Personal Vocabulary, Word Preference Degree, Synonym Graph, File Keyword Vector and so on. Based on this conceptual model, we further put forward an incremental approach about constructing personal synonym vocabulary and synonym graph, and a synonym-based file search method. The experiments prove that the method proposed in this paper shows better recall, precision and performance than traditional desktop search tools, and we get some interesting observations in experiments about personal file manage- ment behaviors.

关 键 词：个人信息管理 文件搜索 访问日志 同义词图 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]