Research on Model Building Method of Accurate Precipitation Forecast
Lin Cheng · Jisuanji fangzhen · 2013
Research the construction of the precipitation forecast model to improve the forecast accuracy.In the precipitation forecast technology,choosing more predictors to construct forecast model,and training to get accurate model by the obtained relationship between the forecast factors,but the nonlinear contact between forecast factors is very difficult to describe because the influence of unstable air.In order to solve this problem,this paper put forward a new precipitation forecast model construction method based on fuzzy clustering algorithm.According to the flow characteristics of air,the fuzzy clustering algorithm was used to analyse the internal directly related characteristics of forecast factors to realize the accurately statement of the contact between forecast factors.Then the initial forecast model was constructed according to the least-squares regression method training,and the precipitation forecast model was acquired.Experiment results show that the fuzzy clustering algorithm can effectively obtain the connection between the forecast factors,and accurately construct and train the forecast model,realizing accurate precipitation forecast.