情报杂志Issue(9)：158-162,5.DOI:10.3969/j.issn.1002-1965.2015.09.030
基于LDA主题特征的微博转发预测
Predicting Retweeting Behavior Based on LDA Topic Features
摘要
Abstract
Retweeting is both the key way for information propagation in microblog and one of the most important issues in information dissemination and public opinion monitoring research. It is significant and helpful to study user retweeting behavior for analyzing informa-tion diffusion,monitoring public opinion and hot extraction. However,most of the researches are focused on the macroscopic stratum of the microblog social network which are difficult to apply on areas such as recommending ads or message and predicting the diffusion path. To predict whether a tweet will be retweeted by a user at the microcosmic level,this paper first builds the microblog features and the user fea-tures based on the analysis of the factors that affect retweeting behavior,and then establishes a predicting retweeting behavior model by combining the latent topic features extracted via using the semantic model of latent Dirichlet allocation with the microblog features and the user features. The experiment results have verified the feasibility of this model based on topic features.关键词
微博转发/主题特征/Latent Dirichlet AllocationKey words
microblog retweeting/topic feature/latent dirichlet allocation分类
社会科学引用本文复制引用
李志清..基于LDA主题特征的微博转发预测[J].情报杂志,2015,(9):158-162,5.基金项目
广州市社会科学规划项目“基于物联网技术的广州智慧社区应用研究”(编号:2012QN06)。 （编号:2012QN06）
