A Study On The Simulation And The Formation Mechanism Of Urban Expansion Based On The Bayesian Network

As the results of the rapid urbanization, city’s environmental problems become worsensharply. It is meaningful to solve the problems appeared in the process of urbanization. It isimportant to do the research to analyze and simulate the characteristics and laws ofurbanization of typical study areas in China. The study of the impact of urban land usedynamic change can provide theoretical support for the decision making and design of theoverall planning in the process of urban construction. The paper takes Jingjinji metropolitanarea economy as the study area, where is the typical area in the process of Chinese urbanconstruction.At present, the study on the land use change and urban land use change is going on andthe scholars build many simulation models to forecast the dynamic change of urban land.These model parameters obtained by logistic regression, artificial neural network, CAtheoretical model, these models are established urban land use change and driving parametersof the relationship between the impact factor, but ignore the driving influence the impact ofthe interaction between the factor.Bayesian network model proposed in this paper through the machine automaticallylearning Bayesian network structure and parameters, to deepen structural relationship betweeneach node variable factor and the parameters of the relationship, by the node variable betweenthe conditional probability distribution table to replace the artificial neural network and othermethods determined by the weight value of each pixel in the data set specific transitionprobabilities, urban land use probability reduce the Jingjinji economic metropolitan areawithin each node variable in determining the model parameters between the complexity of themodel simulation require human involvement Results display using Bayesian network modelsimulated probability ROC test results for analog high probability precision0.9040.975than artificial neural network.By the software of Arcgis, the paper analyzes the structural changes and the conversionof every land use types of Jingjinji economic metropolitan area during2000to2008. It alsoanalyzes the change of the area that the urban construction land invades and occupies otherland use types. At then, reveal the changes and law of urban land use expansion of the studyarea from various angles. On the basis of the above results, analyses the driving factors thatimpact the urban land use expansion of Jingjinji economic metropolitan area. The BNTLCmodel system is applied to the Jingjinji metropolitan economic Metropolitan city land-usechange will establish a case study, the model simulation results show that the accuracy of themetropolitan area based on the Jingjinji economic BNTLC model simulation of urban landuse, after The kappa test a single urban land use simulation accuracy0.934, in order to be ableto determine this article BNTLC model its has good practicality. Finally, this modelsimulation and prediction of urban land use on the Jingjinji economic metropolitan area in2020and forecast urban land use changes in the landscape pattern of expansion and impactanalysis of landscape pattern change in2000-2020number of landscape patches of the townsland within the study area, patch density, the percentage of the index, the largest patch indexhas increased shape index from21.087in2008reduced to20.2372, discrete extent of urbanland use in the region convergence.