Research of public opinion hotspot detection model based on Web big data

In the Web public opinion hotspot analysis, the opinion and emotion in the Web comment play an important role in identifying news topic and emphasis, where few research are carried out. In addition, reports from different data source are processed with the same weight, which may not actually describe the public opinion. This paper propose a hot public opinion discovery model based on quantitative comment and emotion. In this model, Web opinion mining is applied in presence of report vector. By constructing the Web opinion dictionary, the inclination and intensity of Web comment are quantitated by the form of vector. A modified Page-Rank algorism is also proposed to evaluate the source of the Web opinion. Experiment shows that this model can reduce the false detection rate and missing rate of opinion discovery.