A study on DS-SVM-based forecast model from the view of information fusion

作者：Zhu, Weidong; Sun, Yibo; Wu, Yong; Liu, Jingyu

来源：International Journal of u- and e- Service, Science and Technology, 2015, 8(2): 395-404.

DOI：10.14257/ijunesst.2015.8.2.38

• D S evidence theory
• Dempster
• Dempster-Shafer theory
• FORECAST model
• Forecasting modeling
• High-dimensional
• Propensity score
• SVM

Dempster-Shafer Theory is specially advantaged in information fusion, while Support Vector Machine (SVM) can well deal with high-dimensional limited sample data. This Article firstly forecasts the data samples by categories with multiple SVMs, and hence based thereon, fuses the resulting information from multiple SVM models by using DS theory. At the end, Anderson's Iris data set is used to simulate the system of the created DS-SVM model, which shows that the approaches proposed in this Article can not only increase the accuracy rate of the SVM forecasting model, but also improve the propensity scores of the SVM forecasting model.

• 出版日期2015
• 单位合肥工业大学