摘要
[目的/意义]探究社交网络中影响群体情感行为的影响因素,旨在为舆情引导提供参考。[方法/过程]首先,获取微博数据集,进行预处理后,使用结合AP算法及TF-IDF算法的LDA主题模型对用户文本进行聚类,挖掘用户兴趣主题;然后,人工识别结合接口调用对用户文本进行情感标注;最后,利用二元逻辑回归模型对假设的群体情感影响因素进行验证。[结果/结论]性别、主题及活跃度对于群体情感倾向有显著影响;男性群体普遍比女性群体消极;不同主题的情感主旋律不同。该结果对于网络舆情的控制以及个性化新闻推荐具有很好的借鉴作用。
[Purpose/significance]The paper explores the influencing factors of group emotional behavior in social networks in order to provide reference for public opinion guidance.[Method/process]Firstly the paper obtains the microblog data set and after preprocessing uses the LDA topic model combined with AP algorithm and TF-IDF algorithm to cluster user texts to mine user interest topics;then human recognition and interface call are used to annotate the user’s text;finally the binary logistic regression model is used to verify the hypothetical factors affecting group emotion.[Result/conclusion]Gender theme and activity have significant effects on group emotional orientation;The male group is generally more negative than the female group.Different themes have different emotional themes.The results have a good reference for the control of online public opinion and personalized news recommendation.
作者
史伟
王晓飞
Shi Wei;Wang Xiaofei(School of Economics and Management Huzhou University,Huzhou Zhejiang 313000;Schoolof Information Engineering Huzhou University,Huzhou Zhejiang 313000)
出处
《情报探索》
2023年第1期80-86,共7页
Information Research
基金
国家社会科学基金一般项目“重大突发事件中网民情感状态演变规律及引导研究”(项目编号:20BXW013)成果之一。
作者简介
史伟(1981-)男,博士,教授,研究方向为商务智能与情感计算,已发表论文18篇;通讯作者:王晓飞(1994-),女,2020级硕士研究生,研究方向为网络舆情与情感分析。
