Analysis of Weibo Comments Based on SVM and LDA Models

摘要

Nowadays, Sina Weibo is an important way for people to voice their opinions in China. With the development of social media computing, we can find out public opinions of an event by analyzing Weibo comments using computational methods. In this paper, we focus on a hospital event which happened in China. First, we collect comments about the event from Sina Weibo blogs of 4 different kinds of bloggers including traditional media, medical media, Internet celebrity and actor self-media. Then we estimate the sentiment of the comments by making use of Support Vector Machine (SVM) model. Finally, we extract topics from the comments by using the Latent Dirichlet Allocation (LDA) topic model for the purpose of explaining the reason of different sentiment for each blogger.