基于DPCA的社会化问答社区用户生成答案知识聚合与主题发现服务研究  

Research of User-generated-answer Knowledge Aggregation and Topic Discovery Service in Social Q&A Community Based on DPCA

作　　者：陶兴 张向先[1] 郭顺利[2] 

机构地区：[1]吉林大学管理学院,吉林长春130022 [2]曲阜师范大学,山东日照276826

出　　处：《情报理论与实践》2019年第6期94-98,87,共6页Information Studies:Theory & Application

基　　金：国家社会科学基金项目“大数据驱动下学术新媒体知识聚合及创新服务研究”的成果之一,项目编号:18BTQ085

摘　　要：[目的/意义]通过密度峰值聚类算法(DPCA)对社会化问答社区用户生成答案进行知识聚合与主题发现。[方法/过程]利用TextRank方法挖掘用户生成答案中的关键词集合,再利用DPCA对关键词集合进行凝聚,最后进行可视化词云展示。[结果/结论]与传统聚类算法对比,DPCA能获得更加准确的聚类数目,簇类内的关键词更加凝练,知识主题更加明确。文章通过聚合服务,自动化地发现社会化问答社区中的知识主题,进一步提高了社会化问答社区中的用户交流方式和知识服务水平。[局限]用户生成答案缺少标准实验参考集,聚类结果测度有待完善。[Purpose/significance] This paper proposes a method of user-generated-answer knowledge aggregation and topic discovery service in Social Q&A Community based on clustering technology.[Methods/process] TextRank was used to find the text label from user-generated-answer and the Density Peak Clustering Algorithm was used to get the clusters. Then the knowledge clustering result was shown by visual analysis.[Results/conclusion] Compared with the traditional clustering algorithm, DPCA can obtain more accurate clustering number, the key words in the cluster are more concise, and the knowledge theme is more specific. Through the aggregation service, this paper automatically finds the knowledge topic in the social Q&A community, and further improves the user communication way and knowledge service level in the social Q&A community.[Limitations] There is no standard experimental reference set to test user-generated-answer, and the measure of clustering results needs to be improved.

关 键 词：社会化问答社区 知识聚合 主题发现 密度峰值聚类算法 TextRank 

分 类 号：G252[文化科学—图书馆学]