关联规则在领域知识推荐中的应用研究  

Research on the Domain Knowledge Recommendation Based on Association Rules

作　　者：雷雪[1] 侯人华[1] 曾建勋[1] 

机构地区：[1]中国科学技术信息研究所,北京100038

出　　处：《情报理论与实践》2014年第12期67-70,66,共5页Information Studies:Theory & Application

基　　金：国家自然科学基金项目"基于海量数字资源的科研关系网络构建研究"的成果;项目编号:71273251

摘　　要：文章将关联规则挖掘引入共词分析领域。通过试验发现,设置合理的最小支持度和最小置信度阈值,可以有效揭示目标领域的核心二元词组和多元词组间的关联关系。相比传统的共词分析方法,在精确计量词组内部术语间语义关系的强度和方向上具有明显优势。进而提出基于关联规则的信息检索系统知识推荐思路。This paper applies the association rules mining in co-word analysis field. The experiment finds that set a reasonable minimum support and minimum confidence threshold can effectively reveal the relationship between core double phrases and ternary phrases in the target domain. Compared with the traditional co-word analysis,association rules have obvious advantages in precise measurement strength and direction of semantic relationships between terms. Then,the paper puts forward the knowledge recommendation method of information retrieval system based on association rules.

关 键 词：关联规则 共词分析 知识推荐 

分 类 号：TP311.13[自动化与计算机技术—计算机软件与理论]