基于粗糙集的文本分类方法在网络科技资源应用集成环境中的应用  

APPLYING ROUGH SETS BASED METHOD FOR TEXT CATEGORIZATION TO INTEGRATED ENVIRONMENT OF NETWORK TECHNOLOGY RESOURCE APPLICATION

作　　者：侯凡[1] 周明全[2] 耿国华[1] 李杰[1] 

机构地区：[1]西北大学信息科学与技术学院,陕西西安710127 [2]北京师范大学信息科学与技术学院,北京100875

出　　处：《计算机应用与软件》2009年第3期88-91,共4页Computer Applications and Software

基　　金：国家科技基础条件平台建设项目(2005DKA3900)

摘　　要：网络科技资源应用集成环境所汇集到的信息纷繁复杂,使得用户对信息的浏览、检索造成了一定的困难。首先对所有汇集到的信息向量化,然后通过对IF-IDF权重构造函数进行了改良,使其更加适合本项目的实际情况,接着利用粗糙集理论进行属性约简,生成最终的决策表对科技信息进行分类。最终结果证明,提出的分类系统比传统人工分类的效率有较大提高,取得了良好的效果。The information collected by Integrated Environment of Network Technology Resource Application is too complicated to browse and retrieve for users. In this paper it changes all the collected information into vectors, and then by improving the IF-IDF weighting function to make it adapting actual situation of this project better. After that,the concept of rough sets was used to reduce information＇ s attribute,the final decision table was generated to classify the science and technology information. According to the end results, the categorization system proposed in this paper has sharply improved the efficiency of classification comparing with conventional artificial categorization and has good effect.

关 键 词：文本分类 权重函数 粗糙集 属性约简 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程] TH166[自动化与计算机技术—控制科学与工程]