Study On Predicting The Construction Land Of The County

Prediction of construction land is a major research project of the Comprehensive Land Use Planning. Reasonable prediction can meet the needs of socio-economic development and prevent the the disorganized expansion of the construction land. Based on qualitative and quantitative analysis,this paper take Zhao'an as an example to study the prediction models of construction land.Based on the analysis of impact factors of construction land, Dummy Multiple Regression, GM(1,1) and Curve Model were established to predict construction land. Analytic Hierarchy Process(AHP) method was applied to select the best model to improve prediction research and determine reasonable scale of construction land. The main conclusions of this study were as follows:1. With the developing process of urbanization and industrialization, the scale of construction land was in expansion, which expanded faster before 1999. This expansion was obviously slowed down between the year 2000 and 2007 under the macro-control of the land policies from Government. Overall, the increased construction land which up to 1069.9 hectares was mainly used for residential land and transportation land.2. Population, Urban Population, the rate of Urbanization, GDP, per capita GDP, Industrial Output, Investment in Fixed Assets, Industrial Structure, per-construction of GDP and per-construction of Industrial Output were selected from the perspectives of social, economic and land-use efficiency. The results of the Gray Correlation Analysis showed that the correlation degree between population and construction land was the highest,which means that population play an important role in construction land, and the effect of social factors were significantly greater than economic factors.3. According to the requirements of the independence variable for Multiple Regression, Stepwise method was applied to removed the collinearity factors. Then, the population was determined as the independence variable. This chang of construction land showed that the slowing expansion of construction land was related to the Government's land policies. As the result, Dummy variable, which be introduced to qualified the land policies, was used to established the Multiple Regression Model combined population. The result of the model showed that R square of the model was 0.972, and the model passed the T-test and residual test, which means that the model was reasonable. Land system can be considered as a gray system, so GM(1,1) was established based on gray system theory. The accuracy of the model showed that the error was up to the mustard and passed the accuracy test. Considered the trend of construction land was not entirely linear, Quadratic model, Cubic model and Growth model were established. Comparing with the three curve models, Growth model was selected to be a prediction model.4. If the advantages and disadvantages of the Dummy multiple Regression, the GM(1,1), and the Curve Model were considered from the perspectives of theory-scientific, the model-succinct , and accuracy. The AHP method could be used to analyze the weighted sequence of these three models: the Dummy Multiple Regression, the Curve Model, and the GM(1,1). Finally, the Dummy Multiple Regression was selected to be the most suitable model used in the construction land in Zhao'an County, and the area of construction land of this county will up to 6911.8 hectares in 2015.