Internet Public Opinion Hotspot Detection and Analysis Based on Kmeans and SVM Algorithm

Abstract

Rapid progress of network arouses much attention on Internet public opinion, it is important to grasp the internet public opinion in time and understand the trends of their opinion correctly. Text mining plays a fundamental role in categorization and monitoring of internet public opinion, but internet public opinion is much more difficult than pure-text process because of their semi-structured characteristic. To address this issue, we propose a model for internet public opinion hotspot detection and analysis. Due to the text format of internet public opinion, we introduce the traditional vector space model (VSM) to express them, and then use Kmeans algorithm to perform text clustering on a corpus collected from some news website, and use SVM classifier to perform text categorization for new text opinion analysis, the result of the experiment shows that the efficiency and effectiveness of such method.