基于主题的政务微博评论用户画像研究  

Research on User Portrait of Government Microblog Comments Based on Theme

作　　者：王志刚[1] 邱长波[1] Wang Zhigang;Qiu Changbo(School of Management,Jilin University,Changchun 130022)

机构地区：[1]吉林大学管理学院,长春130022

出　　处：《情报杂志》2022年第3期159-165,共7页Journal of Intelligence

摘　　要：[研究目的]构建政务微博评论用户的画像,直观地展现各群体的典型人物特征,有助于了解各群体用户特征和行为之间的关系,对政府改善微博的管理策略有着重要的意义。[研究方法]使用LDA模型挖掘出用户评论中的偏好主题,并得到每个用户偏好主题的概率分布向量,据此计算出用户的相似度,采用多维标度法实现用户群体的分类,最后利用Logistic回归模型提取出各群体用户的典型特征。[研究结论]政府可以根据用户群体的典型特征,向用户精准推送相关的主题内容,增强政民沟通的效率;不同特点的用户群体的评论会发挥出不同的作用,政府应该采取不同的处理方式,增强公众参与的效果。[Research purpose]Constructing the portrait of government microblog comment users,intuitively showing the typical character characteristics of various groups,helps to understand the relationship between the characteristics and behaviors of various groups of users,and is of great significance for the government to improve the management strategy of microblog.[Research method]The LDA model is used to mine the preference topics in user comments,and the probability distribution vector of each user preference topic is obtained.Based on this,the similarity of users is calculated,and the multi-dimensional scaling method is used to classify user groups.Finally,the typical characteristics of users in each group are extracted by logistic regression model.[Research conclusion] The government can accurately push relevant theme content to users according to the typical characteristics of user groups,so as to enhance the efficiency of government people communication;the comments of user groups with different characteristics will play different roles;the government should adopt different ways to enhance the effect of public participation.

关 键 词：政务微博 政务微博用户 评论用户 用户画像 用户偏好 主题偏好 群体分类 

分 类 号：G206[文化科学—传播学]