A novel rainfall forecast model based on integrated non-linear attributes selection method and support vector regression

Rainfall forecast has been a popular research topic. The precise rainfall prediction can not only assist water management in water scarcity regions, but also can warn or relief the effects of excess or insufficient rainfall. As a result of the advancement in information technology, the current prediction methods are more diverse and sophisticated, however spending a huge amount of resources, time, and costs, the forecast outcomes are still very uncertain. Therefore, this study proposed a novel rainfall forecast model, which combined proposed integrated non-linear attributes selection method with support vector regression (SVR) to enhance the forecast performance. Firstly take advantage of the proposed integrated non-linear attributes selection method to find the important attributes that affect the rainfall in Taiwan mountain areas, and then input the data of selected attributes into SVR to train rainfall forecast model. In order to assess the prediction performance of the proposed model, this study practically collected the rainfall data from 2005 to 2014 in Taiwan Mountains monitoring stations, and compares the proposed model with the listing models. Experimental results show that the proposed model outperforms the listing models in RMSE.