Dynamic Prediction of Red Date Yield Based on 3 Kinds of Regression Models

Based on yield data of red date and corresponding meteorological element data of Zhongwei City in Ningxia in 1981-2005,the relationships between meteorological factors and yield in return-to-green stage,flowering stage and fruit color-changing stage of red date were analyzed by applying entering regression model,stepwise regression model and backward regression model.And the calculation results of models classified on the basis of poor and rich harvest grade of yield and the excellence degree of model was assessed through mathematical expectation.The results showed that the accurate rate of predicting poor and rich harvest degree of yield in current year by applying stepwise regression model in return-to-green stage,backward regression model in flowering stage and entering regression model in fruit color-changing stage was higher.