Climate Prediction by SVM Based on Initial Conditions
Wang Deji, Bo Xu, Zhang Faquan, Li Jianting, Guangcai Li, Sun Bingyu · 2009
The climate model is the crucial factor for agriculture. However, the climate variables, which were strongly corrupted by noises or fluctuations, are complicated process and can not be reconstructed by a common method. In the paper, we adapt the SVM to predict it. Specifically, we incorporate the initial condition on climate variables to the training of SVM. The numerical results show the effectiveness and efficiency of the approach.