摘要
针对检索系统快速优化问题,提出了一种自动构建用户概念空间的方法。概念空间的构建是实现语义检索的关键,为语义检索提供知识源,传统的面向通用领域的概念空间过于庞大,实现起来非常困难。面向每个用户时,给出了一种构建用户私有概念空间的方法,私有概念空间是根据用户检索文档历史记录自动生成的。在构建用户概念空间的过程中,引入粒计算的思想,用二进制粒表示频繁特征项,并通过粒之间的运算挖掘频繁2-项集,最后计算特征项之间的语义关联权,进行了仿真。仿真结果表明,方法具有较高的效率。
A method for building user concept space automatically is presented in this paper.The construction of concept space is the key to realize the semantic retrieval,which provides knowledge resource for semantic retrieval,.However,the traditional concept space of general fields is too large,and is very difficult to be constructed.Therefore,this paper proposes a method to build private concept space for each user,which is generated automatically by user＇s retrieval documents history,moreover,it also can help users get personalized service.In the process of building user concept space,the thought of granule computing is introduced to this paper,binary granules are used to represent the frequent feature items,and the frequent 2-itemsets are mined by calculation among granules.Finally,the semantic association weight between feature items are computed by the calculation expression,then the user concept space can be obtained.Simulation experiment results show that this method proposed in this paper has better time efficiency than the classic Apriori algorithm.
出处
《计算机仿真》
CSCD
北大核心
2011年第6期223-226,共4页
Computer Simulation
基金
江西省2007年科技攻关计划项目(2007CC02000)
关键词
二进制粒
用户概念空间
频繁集
语义关联权重
Binary granule
User concept space
Frequent item set
Weight of semantic relationship
作者简介
谭德坤（1973-），男（汉族），重庆开县人，讲师，博士研究生，主要研究方向：文本挖掘，智能信息检索与粒计算。
