基于机器学习与空间大数据的养老设施选址应用研究OA

With the acceleration of population aging,the current elderly care service system faces dual pressures of resource scarcity and imbalanced regional allocation,which is particularly prominent under urban-rural gradient disparities.This paper took the main urban area of Nanning as an empirical case,integrates the geospatial points of interest(POI)obtained from Amap API and the data of the Seventh National Population Census,and built a site selection optimization model based on the decision-making tree.By analyzing the distribution characteristics of the facilities across 26 509 spatial grids,the paper simulated 1 993 potential locations for the elderly-care facilities.These were further refined to 787 priority planning zones based on the neighborhood aging rates and population density,and finally,41 locations for new elderly care facilities were planned and identified through a maximum coverage model.It is found that:①the predicted locations of the model demonstrated 89.53%consistency with the existing facilities,and the predicted locations are largely overlapping with the government-released planning sites for the elderly-care facilities,confirming its reliability.②The re-screened sites planned in priority and the newly-added elderly-care facilities of the grid are predominantly located in sub-districts(townships)with relatively severe aging populations,such as Fujian Garden Sub-district and Nanyang Town.These predictive outcomes align with the actual distribution of the elderly residents in Nanning City.This paper demonstrates the effective application of big data and machine learning technologies in the site selection of the elderly care facilities,providing a reference for refined and scientific urban planning.