基于社交关系的微博主题情感挖掘

发布时间：2019-04-08 08:40

【摘要】：微博情感分析是社交媒体挖掘中的重要任务之一,在个性化推荐、舆情分析等方面具有重要的理论和应用价值.挖掘性能良好且可同步进行文档主题分析与情感分析的主题情感模型,近年来在以微博为代表的社交媒体情感分析中备受关注.然而,绝大多数现有主题情感模型都只简单地假设不同微博的情感极性是互相独立的,这与微博生态的现实状况不相一致,从而导致这些模型无法对用户的真实情感进行有效建模.基于此,综合考虑了微博用户相互关联的事实,提出了基于LDA和微博用户关系的主题情感模型SRTSM(social relation topic sentiment model).该模型在LDA中加入情感层与微博用户关系参数,利用微博用户关系与微博主题学习微博的情感极性.针对新浪微博真实数据集上的大量实验结果表明:与代表性算法JST,Sentiment-LDA及DPLDA相比较,SRTSM模型能够对用户真实情感与讨论主题进行更加有效的分析建模.
[Abstract]:Weibo emotional analysis is one of the important tasks in social media mining. It has important theoretical and applied value in personalized recommendation and public opinion analysis. The thematic affective model, which has good performance and can be used synchronously for document thematic analysis and affective analysis, has attracted much attention in recent years in social media affective analysis represented by Weibo. However, most of the existing thematic affective models simply assume that the emotional polarity of different Weibo is independent of each other, which is not consistent with the reality of Weibo ecology. As a result, these models can not effectively model the real feelings of users. Based on the fact that Weibo users are related to each other, a thematic affective model based on the user relationship between SRTSM (social relation topic sentiment model). And SRTSM (social relation topic sentiment model). Is proposed. In this model, emotional layer and Weibo user relationship parameters are added to the model, and Weibo user relationship and Weibo theme are used to learn the emotional polarity of Weibo. A large number of experimental results on the real data set of Sina Weibo show that compared with the representative algorithms JST,Sentiment-LDA and DPLDA, the SRTSM model can be used to analyze and model the real feelings of users and the topic of discussion more effectively.
【作者单位】：东北大学计算机科学与工程学院;福建师范大学软件学院;福建省公共服务大数据挖掘与应用工程技术研究中心;医学影像计算教育部重点实验室(东北大学);广西财经学院信息与统计学院;广西师范学院计算机与信息工程学院;Grasslands
【基金】：国家重点基础研究发展计划(973)(2012CB316201) 国家自然科学基金(61433008,61363009,61363037) 福建省教育厅K类科技项目(JK2016007)~~
【分类号】：TP391.1

【相似文献】