图书情报工作 ›› 2018, Vol. 62 ›› Issue (20): 122-132.DOI: 10.13266/j.issn.0252-3116.2018.20.014

• 知识组织 • 上一篇    下一篇

知乎平台用户影响力分析与关键意见领袖挖掘

郭博1, 许昊迪2, 雷水旺3   

  1. 1. 珠海市魅族科技有限公司北京分公司 北京 100872;
    2. 香港科技大学 香港 999077;
    3. 徐州工业职业技术学院图书馆 徐州 221140
  • 收稿日期:2018-03-28 修回日期:2018-07-19 出版日期:2018-10-20 发布日期:2018-10-20
  • 作者简介:郭博(ORCID:0000-0002-4053-749X),数据专家,高级工程师,博士研究生,E-mail:guobo01@126.com;许昊迪(ORCID:0000-0001-6016-9084),硕士研究生;雷水旺(ORCID:0000-0002-7120-4460),馆员,硕士。

Analysis of User Influence and Identification of Key Opinion Leaders Based on Zhihu Platform

Guo Bo1, Xu Haodi2, Lei Shuiwang3   

  1. 1. Meizu Telecom Equipment Co., Ltd. Beijing 100872;
    2. The Hong Kong University of Science and Technology, Hong Kong 999077;
    3. Xuzhou College of Industrial Technology Library, Xuzhou 221140
  • Received:2018-03-28 Revised:2018-07-19 Online:2018-10-20 Published:2018-10-20

摘要: [目的/意义]随着互联网技术的快速发展,知乎平台逐渐成为一个热议社会公众话题以及分享知识、经验的载体。因此,分析知乎平台中关键用户的影响力和挖掘其中的关键意见领袖在研究社交网络信息传播途径的过程中起到非常重要的作用。[方法/过程]通过提出改进的PageRank算法和HITS算法,分别基于知乎用户社交网络、问答网络构建用户影响力挖掘模型,能够准确、客观地识别出其中的关键用户及意见领袖。[结果/结论]实验结果表明,提出的PageRank算法和HITS算法能够有效挖掘出知乎平台中具有较为突出特性的关键意见领袖,并且算法的收敛速度较快,具有可复用性和迁移性。通过对知乎平台用户数据集进行处理和有效分析,成功建立用户影响力和关键意见领袖挖掘模型;同时,在具体话题上进行验证。因此,可以推断该模型有巨大应用价值和商业化推广前景。

关键词: 知乎, 用户影响力, 关键意见领袖, PageRank算法, HITS算法

Abstract: [Purpose/significance] With the rapid development of network technology, the platform of Zhihu has become a significant carrier to discuss social public topics and share knowledge as well as specified experience. Therefore, it is of importance for studying the communication channels of social network information to investigate the influence of key users and dig out the key opinion leaders in the Zhihu platform.[Method/process] By the means of improved PageRank and HITS algorithms, this study constructed a model for evaluating user influence based on the social network and question answering network of Zhihu platform, and identified the key users and opinion leaders accurately and objectively.[Result/conclusion] The experimental results show that PageRank and HITS algorithms in this paper could effectively extract several key opinion leaders with prominent features in Zhihu platform, the speed of the convergence is fast and with high reusability and mobility. By processing and analyzing the user data set of Zhihu platform, we successfully build a model for evaluating the user influence and mining key opinion leaders. Along with the verification of specified topics, it can be inferred that this model has enormous application value and commercial promotion prospect.

Key words: Zhihu, user influence, key opinion leader, PageRank algorithm, HITS algorithm

中图分类号: