An effective framework based on nerve cell theory and topic aware feature statistics for Internet public opinion hotspot forecasting

Internet public opinion has become a popular topic of web data mining. Many researches aim at hot topic detection to help understand what is happening, but very a few works focus on forecasting public opinion hotspot to assistant learn what will happen, although this also has great significance. A framework for Internet public opinion hotspot forecasting using nerve cell theory and topic aware feature statistics is proposed in this paper. It takes the neuron´s stimulus-response theory, multipass convey theory into the predicting of hotspots, and integrates topic detection and feature statistics techniques to achieve the goal of forecasting. Experimental results validated that the proposed framework is effective.