Using Data Mining Method Combined Co-integration Theory To Predict Precipitation

Along with the increasing scarcity of water resource and seriousness of drought,the rational planning and utilization of water resources has become one of the hottest topic.Precipitation,as a renewable fresh water resources,its forecast becomes the key of the research problem.This paper proposes a novel method-data mining based on co-integration theory to predict the precipitation.This hybrid model mainly uses co-integration theory to introduce the relevant information and uses the machine learning to depict the nonlinear relationship.Compared with the traditional machine learning algorithms,the co-integration theory could significantly improve the accuracy of prediction.Therefore,the hybrid model plays an important role in the utilization of agricultural water and the work of drought relief.