An optimized combination model for construction land increasing trend forecasting

The trend prediction of urban construction land increasing offers a scientific basis for land use decision-making. The data analysis models such as the exponential model, the logistic model, the gray system model and the binary linear regression model are generally used in the trend prediction of urban construction land increasing. Due to the different requirement for data and various fitting models, the prediction results based on above models sometimes have some differences and can't be selected rationally when the difference is larger. The optimized combination model based on non-linear programming, taking the constrained condition of minimal error into account, can synthetically analyze and compare above mentioned single prediction model and reduce the error of construction land prediction. Taking Wuhan as an example, the exponential model, the logistic model, the gray system model and the binary linear regression model are used in this paper to forecast the demand for construction land of Wuhan in the year 2010, 2015 and 2020. Based on this, confirming the weight coefficient of the four prediction models in optimized composite model, optimized prediction result can be obtained. The results indicate that optimized composite model can simulate the trend of construction land increasing much better.