基于类别参与度的社区问答专家发现方法  

Category participation-based approach to find experts for community question answer services

作　　者：林鸿飞[1] 王健[1] 熊大平[1] 刘晓鸣[1] 

机构地区：[1]大连理工大学计算机科学与技术学院,辽宁大连116024

出　　处：《计算机工程与设计》2014年第1期333-338,共6页Computer Engineering and Design

基　　金：国家863高科技研究发展计划基金项目(2006AA01Z151);国家自然科学基金项目(60973068;61277370);辽宁省自然科学基金项目(201202031);教育部留学回国人员科研启动基金;高等学校博士学科点专项科研基金资助课题基金项目(20090041110002)

摘　　要：为了提高社区问答系统的服务质量和效率,使得提问用户尽快得到具权威性的满意回答,研究了专家发现问题,提出了一种基于用户类别参与度的专家发现方法。计算用户在每一个类别的初步专家得分,以及两两类别间的相似度,得到用户对每一个类别的参与度,线性综合用户在本类别的初步专家得分和其它相近类别的参与度得分,即为用户在本类别的最终专家得分。实验是在Yahoo!Answers上抽取的真实标注数据集上进行的。实验结果表明,该方法有效且可行。To improve the service quality and efficiency of community-based question answering systems, and to satisfy the users as soon as possible by making the users＇ questions answered authoritatively, the problem of identifying and discerning expert users is studied. To solve the problem, a method is presented which is based on category participation to find experts in commu nity-based question answering services. Firstly, the original expert score of the user in each category is calculated, as well as the similarity between the categories. Secondly, the category participation of the user for every category is obtained. Finally, the o verall expert score of the category is a linear combination of original expert score in that category and participation score of other similar categories. The experiments are conducted on the question-answer threads of the Yahoo! Answers, and the final results show that this method has a good performance.

关 键 词：社区问答 专家发现 链接分析 类别参与度 相似度计算 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]