EMPIRICAL EVIDENCE FOR STEADY-STATE DISTRIBUTION PREDICTION OF URBAN POPULATION BASED ON DECISION TREE MODELLING:TAKING QUANZHOU BAY AREA AS AN EXAMPLE
An important task of urban planning is to allocate urban space resources based on the predicted spatiotemporal variation characteristics of urban population.Urban population moves rapidly in different spaces and has the characteristics of non-equilibrium distribution.The spatial and temporal changes of the uneven distribution of the population have a certain regularity,that is,the steady-state distribution.Based on LBS data and multi-source big data of the built environment,this paper selects 4 categories and 36 sub-categories,and adopts the intelligent decision tree model to construct a prediction model for the steady-state distribution of urban population.As for the empirical findings in Quanzhou,the accuracy of the prediction model is 84.32％,which provides a solid basis for its reliability.Its internal theoretical basis is the steady-state spatiotemporal activity law of urban population in space,which can support decision-making for future urban planning.
spatial and temporal distribution of populationmulti-scenario predictionplanning evaluationrandom decision forests
NETL
NSTL
万方数据
