Study of Prognostic Method for Distributive Patterns of Summer Rainfall Based on Fuzzy Support Vector Machine

Abstract

Business practices over the years show that three types of rainfall patterns basically reflect the characteristics of summer rainfall in eastern China, and have great practical value for business forecasting. Therefore, in this paper, on the basis of one-versus-one support vector machine multi-classification algorithm, combining with the fuzzy support vector machine, a new model is put forward, and we apply it to the prognosis of summer rainfall pattern. The experiments show that the model has better results than the ordinary one-versus-one SVM multi-classification method and the traditional statistical method on the prognosis experiments of summer rainfall pattern.