Simulation Of LUCC Based On Neural Network And Decision Tree Model

Valuable land resources as a strategic resource,is an important basis for carrying urban economic development,but unreasonable land use structure became the shackles on urban economic development.How scientific,rational and effective use of existing land resources gradually become an important issue for national development.Land Use/Cover Change(LUCC)may demonstrate its spatial distribution of land resources and future trends in the process of urban development,so as to provide a scientific basis for the management and sustainable development policies of urbanization of land resources.The paper is divided into an introduction,body,conclusion of three parts.The introduction highlights the background and significance as well as domestic and foreign research progress in on land use change and simulation,and identify research framework,relevant research methods and technical route based on the content of the selected study.The body the introduction of the principle of land use change,qualitative research on Shiyan City land use change trend,learn that land use change in trend is mainly the expansion of construction land and land transportation,others including the reduction of farmland,forest,water and other land use during 2009-2015,new construction land and transportation extending along the lines of the center of the city limits in the surrounding towns to the urban northeast of Maojian and Bailang Economic Development zone;Reduction of construction land conservation policy implications about the collection of land,mainly scattered in outlying towns,smaller scale.Land use change has important implications for the population,distance to the center city,distance to the construction land,distance to the transportation,slope,elevation and distance to water factors for quantitative analysis,and according to the value to get their location impact analysis map.Then quantitative analysis land use within the scope of Shiyan City,in 2009,2012 and 2015,and the land use changes during 2009—2015,get the transfer matrix of land use change each time,master Shiyan whole land resources structure and land use change trends.The paper on the basis of carding,drawing scholars on land use change related research results were used neural network model and decision tree model to simulate land use change.Firstly,the basic principle of neural network model was constructed and the use of neural network simulation by the combined effects of multiple factors of land use change,Shiyan City by land use type and the driving factors of the relationship between learning and training,continue to adjust its parameters to verify the accuracy of up to meet the accuracy requirements,will eventually trained neural network model to simulate land use change;then introduces the principle of the decision tree,the land will change into analog special use decision tree classification,according to population,neighborhood analysis,slope,elevation and other factors determine the decision to establish the conditions and land use suitability evaluation set,test and simulation accuracy parameters continuously modify,adapt the decision tree model changes in the relationship between the factors and land use.Finally,Decision tree model simulates meet the accuracy requirements of the land use change;neural networks and decision trees have advantages and disadvantages,compared to neural networks,decision trees model by human interference and policy impact is relatively large,more suitable for simulation of construction land and land transportation the combination of both analog Shiyan City land use change.Conclusions,summarized and analyzed the results of this study and pointed out shortcomings of this study for future in-depth exploration.