Abstract:

Prediction of urban builtup area is a core issue in urban studies There is always a complex nonlinear relationship between urban builtup area and urban economy and society It is difficult to accurately predict urban builtup area by using traditional methods and models such as linear regression，time series analysis，gray system theory and BP neural network As a new artificial neural network model，RBF neural network has some advantages of fast learning， easily getting in the local minimum and approximating any arbitrary accuracy of the global nonlinear relationship Therefore，RBF neural network can overcome some shortcomings of BP neural network and show a ability to handle complex nonlinear system Currently，RBF neural network is one of the most accepted prediction methods Taking the prediction of builtup area in Hefei City from 1997 to 2007 as a research sample，this paper established a prediction model based on RBF neural network from five impact indexes including GDP，financial income，total fixed asset investment，nonagricultural population and average wage of workers As a comparison，this paper also used BP neural network，simple linear regression （SLR） model and multiple linear regression （MLR） model to predict The results indicate that the means of prediction residuals and relative errors of RBF neural network were only 04027 km2 and 029% which were the minimum values in the prediction results of the four models； the means of prediction residuals of BP neural network，SLR model and MLR model were 35794 km2，68531 km2 and 36668 km2 respectively； the means of prediction relative errors of BP neural network，SLR model and MLR model were 208%，457% and 238% respectively  The residuals of RBF neural network in each year were the smallest except for 1999； the residuals of SLR model in each year were the largest except for 1999，2004 and 2007 From 2004 to 2005，the builtup area of Hefei was mutated, which led to large errors （＞8 km2） predicted by BP neural network，SLR model and MLR model At the same time，RBF neural network still had a higher prediction accuracy which was less than the mean of prediction residuals （04027 km2） of RBF neural network The prediction results clearly show that the prediction accuracy of RBF neural network is the best，BP neural network is the second，MLR model is the third and SLR model is the worst Compared with BP neural network and linear regression models，RBF neural network can resolve the nonlinear relationships more accurately in complex systems，which makes RBF neural network a higher prediction accuracy in the complex nonlinear conditions Hence RBF neural network can provide a new idea and method for the prediction of urban builtup area Moreover，it can provide scientific basis for decision making in urban land use and planning.