Abstract
Revealing the spatial distribution and regional differences of red blood cell distribution width (RDW) reference values in the healthy Chinese population is important for the monitoring, diagnosis and prevention of the top chronic diseases in the disease spectrum in China. Medical indicator data were matched with geographical data and environmental pollution data to predict RDW reference values for the whole region of China using ridge regression analysis, support vector machine and BP neural network. Taylor diagrams and assignment methods were used to select the optimal model, and the spatial distribution was mapped according to the prediction results of the optimal model, and the intervals of RDW reference values for different age groups of healthy people in China were redefined. It was suggested that the reference values of RDW in the young and middle-aged groups in China were divided into low-value areas in the north (12.31,12.80) and high-value areas in the south (12.80,13.18); the reference values of RDW in the elderly group were divided into low-value areas in the north (11.28,12.60) and high-value areas in the south (12.60,13.91).In turn, it can provide a more accurate reference basis in the monitoring, diagnosis and prevention of chronic diseases at the top of the disease spectrum.