支持用户偏好查询的领域概念图模型  

Domain concept graph model supporting user preference query

作　　者：高志君[1] 郑俊生[1] 安敬民 GAO Zhi-jun;ZHENG Jun-sheng;AN Jing-min(School of Computer,Dalian Neusoft University of Information,Dalian 116023,China;School of Information Science and Technology,Dalian Maritime University,Dalian 116026,China)

机构地区：[1]大连东软信息学院计算机学院,辽宁大连116023 [2]大连海事大学信息科学技术学院,辽宁大连116026

出　　处：《计算机工程与设计》2022年第3期744-750,共7页Computer Engineering and Design

基　　金：国家自然科学基金项目(61602075);辽宁省自然科学基金项目(20180550940)。

摘　　要：针对目前的领域概念查询聚类方法中未见考虑用户偏好,提出一种支持用户偏好查询的领域概念图模型。该图模型主要包括两部分:基于概念本身考虑,利用综合语义相似度计算方法构建概念的语义关系图;基于用户查询偏好考虑,采用改进的互信息计算用户生成数据间隐含的查询偏好,将其结果用于补全领域概念的语义关系图。这一处理过程使得原有领域概念的语义关系图得到了有益的补充,满足了用户的偏好查询。经实验验证,该算法较现有方法,查准率、查全率以及F-measure值均有所提高且响应时间得到了降低。In view of the fact that user preference has not been considered in current domain concept query-clustering methods,a domain concept graph model supporting user preference query was proposed.The model mainly included two parts,considering the concepts themselves,the concept semantic relation graph was constructed using the comprehensive semantic similarity calculation method.Based on the consideration of users’query preferences,the mutual information was used to calculate the implicit query preferences via user generated data,which were used to complete the domain concept semantic relation graph.This process makes the semantic relation graph of the original domain concept get beneficial completion and satisfy the users’preference queries.Results of experiments show that the precision,recall and F-measure of the model are greater,and the running time is reduced,compared with that of other existing methods.

关 键 词：领域概念图 概念聚类 用户偏好 互信息 用户生成数据 语义相似度 

分 类 号：TP391.1[自动化与计算机技术—计算机应用技术]