A geographically weighted model of the regression between grain production and typical factors for the Yellow River Delta

The study of grain production is one of the crucial areas for grain security. In this paper, the quantile results for grain production reveal that the first-quantile county units include Kenli country, Lijin country, Hekou District and Dongying District. The second-quantile, the third-quantile and the fourth-quantile country units include five country units each. Moreover, a model of geographically weighted regression (GWR) between grain production and typical factors was used to explore the spatial adjacency relationship among them; grain production was regarded as the dependent variable, and some typical factors, namely the grain sowing area, the efficient irrigation area, and agriculture machinery, were thought of as the independent variable based on 19 county level areas of the Yellow River Delta in 2007. The GWR was established for predicting the grain production at the county level, the grain production of 14 random counties was regarded as the training data, and the grain production in Changyi country, Huimin country, Gaoqing country, Bincheng District and Lijin country was regarded as the test data. The GWR results showed that no underprediction and overprediction of values for the grain production arose in the study. The adjustment R2R2 value of the grain production in the GWR model is 0.988, standardized values of the residuals including five country units are positive, the predicted value and the observed value for the grain production of 19 county units showed a good fit, and the established GWR model can be used to carry out the prediction of grain production in Yellow River Delta.