基于情感分析的新浪微博争议度分析

随着新浪微博的广泛使用,新浪微博传播的内容涵盖了各种领域的信息,涉及范围越来越广;同时,当用户浏览消息时,不仅能够表达自己的看法,同时能够看到别人的观点。因此,当用户无法判断问题的客观性时,通常会利用其他公众的反馈信息进行评估。为了充分利用公众反馈信息,提出了一种描述公众反馈信息的社会属性——争议度。给出了争议度的概念并构建了争议度计算模型,该模型将公众对微博的反应(即表态、评论、转发等行为)作为争议度的影响因素,通过情感分析等技术计算微博消息的争议度。实验结果对照人工标注结果,模型准确率达到93%,有效率达84%,证明了该模型的可行性。

Sina Microblog disputed-level analysisbased on sentiment analysis

With the popular use of microblogs,the contents of Sina Microblog cover information of various fields,and the rang becomes increasingly wider.As a result,when users of the microblog browse news,they may feel confused concerning what they can believe.So they probably turn to others’ feedback.We propose a concept of disputed-level,a new social property of microblogs,which describes the feedback from public.We define the content of disputed-level and build a disputed-level calculation model.In the model,the influencing factors of disputed-level include comments and forwording.We use the sentiment analysis to calculate the disputed-level.The accuracy rate of the model in the experiment can reach 93%,the efficiency 84%,which prove the feasibility of the model.