Temperature Forecast Model Based on Support Vector Machine Method

Based on the " Function Estimation" and "Cross Validation 1" of the CMSVM2. 0, the temperature data of Hainan from 1970 to 2014 were used to construct the forecasting model of regression method of SVM, and the simulation experiments were performed. The results indicated that the CMSVM2. 0 has good forecasting ability for short-term temperature forecast of Hainan, and the prediction effects of the "Cross Validation 1" is higher than that of the general prediction, especially, in the winter, summer and autumn. Additionally, the prediction effects of SVM for the west, north and south of Hainan are better than that for the east and central.