图书情报工作 ›› 2020, Vol. 64 ›› Issue (18): 76-88.DOI: 10.13266/j.issn.0252-3116.2020.18.009

• 情报研究 • 上一篇    下一篇

社会化问答社区答题者发现研究

潘梦雅, 沈旺, 代旺, 刘嘉宇   

  1. 吉林大学管理学院 长春 130022
  • 收稿日期:2020-04-26 修回日期:2020-06-16 出版日期:2020-09-20 发布日期:2020-09-20
  • 作者简介:潘梦雅(ORCID:0000-0002-0319-626X),硕士研究生,E-mail:pmy156@126.com;沈旺(ORCID:0000-0002-8933-5653),副教授;代旺(ORCID:0000-0001-7168-7776),硕士研究生;刘嘉宇(ORCID:0000-0002-2317-8157),硕士研究生。
  • 基金资助:
    本文系国家自然科学基金项目"基于图模型的多源异构在线产品评论数据融合与知识发现研究"(项目编号:71974075)研究成果之一。

Social Question Answering Community Respondent Discovery Research

Pan Mengya, Shen Wang, Dai Wang, Liu JiaYu   

  1. Management School of Jilin University, Changchun 130022
  • Received:2020-04-26 Revised:2020-06-16 Online:2020-09-20 Published:2020-09-20

摘要: [目的/意义] 识别社会化问答社区中回答可能性高的专业答题者,可缩短提问用户得到满意答案的等待时间,促进用户间的知识共享,助力社会化问答社区的持续健康发展。[方法/过程] 基于社会资本理论及动机理论,对用户答题动因进行分析,结合专家发现研究提出测量指标,构建研究模型,以知乎社区为研究实例,借助Python语言对实验数据进行特征值提取、打标签等数据处理,研究运用逻辑回归模型、随机森林、XGBoost3种常用的机器学习分类模型进行训练及预测。[结果/结论] 与PageRank、HITS算法对比验证本文方法的有效性及优越性,本研究为同类平台如健康社区的问题推送、专家识别以及推荐模型的课题研究提供一定的参考。

关键词: 社会化问答社区, 专家发现, 社会资本理论, 动机理论, 机器学习

Abstract: [Purpose/significance] Identifing the professional answerers with high probality in the social Q&A community can shorten the waiting time for users who ask questions to get satisfactory answers, promote knowledge sharing among users, and contribute to the sustainable and healthy development of the social Q&A community.[Method/process] Based on the social capital theory and motivation theory, this paper analyzed the motivation of users' answering questions, combined the expert discovery research to propose measurement indicators, and built a research model, then took Zhihu as a research example, and used Python to extract the eigenvalues and label of experimental data. Three common machine learning classification models, logistic regression model, random forest model and XGBoost model were used for training and prediction.[Result/conclusion] Compared with PageRank and HITS algorithms, the effectiveness and superiority of the method proposed by this paper have been verified. And this paper has provided a certain reference for the topic research of similar platforms such as healthy community problem push, expert identification and recommendation models.

Key words: social Q&A community, expert finding, social capital theory, motivation theory, machine learning

中图分类号: