利用灰色理论数据预测及MATLAB代码（数学建模）.pdf

Building feasible models It recognizes that there exist a few models to estimate of future performance using currently-available data. For these different methods, each model has some advantages or disadvantages. According to the existing forecasting theory, two models can be offered here: polynomial fitting model During the engineering and scientific research, people usually represent the relationship between two different variables and the variation tendency with the aid of smooth curve on the condition that several reliable data are known. It is supposed that all the original data has several noisy records, as the result ,the function we get only presents the variation trend. Fitted curve is not required to pass all the known data. We use Minimum variance function to assure that the global error during the fitting progress is the minimum. Grey Theory Prediction Model Within the grey system, some of the data are known and the others are unknown. In addition, their relationships are uncertain. The coral of the model is to find regular pattern after the Handling Processing of the original data. Comparison of the two model: We choose the PH of the Suzhou Xishan from week 25 to 34 as the original data. Suzhou Xishan WEEK PH 25 7.56 26 7.26 27 7.31 28 7.47 29 7.6 30 7.41 31 7.55 32 8.0 33 7.55 From: Chinas Ministry of Environmental Protection (MEP) 1.grey system according to this, after the Accumulated Generating Operation to the original data, new reco