The Population Prediction Based On Gray System

Gray system theory is a new method to research on the non-determinism with little data. Studying on the non-determinism system that has characteristic of some information having known while others remain unknown, gray system theory extracts valuable information to describe and effectively through the production and development of some information having known. And gray forecast finds and grasps system tendency to make scientific quantitative forecast of system’s future condition by processing initial date and establishing gray model. The thesis researchers on gray forecast applied in population prediction.This paper improves the accuracy of the gray prediction model in two aspects. First, improve the smoothness of the original data sequence. In this section, we use sin(x) to this transformation. Secondly, improve the mechanism of GM(1,1) model. During this part, we use metabolic GM(1,1) model, new initial value GM(1,1) mode, and improved discrete gray model. The population modeling results show that these three models can improve prediction accuracy.Finally, there are some shortcomings of a single model prediction methods will, from the perspective of data mining, we use the BP neural network to combine all the improved gray model of the above forecasts, and examples prove this model feasibility.