원문정보
초록
영어
In order to enhance prediction precision of online public opinion, it put forward a kind of online public opinion prediction model (PSR-SVR) with the combination of chaos theory and support vector regression. First of all, the original data of online public opinion were obtained throughout topic segmentation, hotspot extraction, and data aggregate. Then, time sequence of online public opinion was reconstructed throughout phase-space reconstruction. Finally, the reconstructed time sequence of online public opinion was input support vector regression for modeling and prediction, and then it was compared with other online public opinion prediction model by experiment. The result shows that compared with the contrast model, PSR-SVR improves the prediction precision and reliability of online public opinion, and the prediction results have certain practical value.
목차
1. Introduction
2. Phase Space Reconstruction and Support Vector Regression
2.1. Phase Space Reconstruction
2.2. Support Vector Regression
3. PSR-SVR Online Public Opinion Trend Prediction Model
4. Simulation Experiment
4.1. Data Source
4.2. Contrast Model and Evaluation Standard
4.3. Pretreatment of Online Public Opinion
4.4. Phase Space Reconstruction of Online Public Opinion Data
5. Conclusion
Reference