Prediction System of Public Opinion in Universities Based on Sentiment Analysis
Designed a prediction system of public opinion in universities based on sentiment analysis.By comparing sentiment dictionaries,the system extracted data of emotion features and data of thematic features from comments,integrated temporal and popularity data,and established an emotional feature model.A loss function was used to correct the model,and SVM was used to predict the potential for public opinion outbreaks.
sentiment analysisfeature extractionmodel of emotional featuresupport vector machineprediction of public opinion
NETL
NSTL
万方数据
