Study on the Site Selection Model of Elderly Care Facilities in Beijing

ABSTRACT : This study takes Beijing as the research area, integrates POI data, ' seventh census ' demographic data and spatial distribution characteristics of pension facilities, and uses machine learning algorithm to simulate and predict the suitable location of pension facilities. The study first screened the areas that meet the site selection conditions from 67,739 grids in the city, and then screened again according to the aging degree of the street, and finally determined 551 key points that need to be arranged for the elderly care facilities, and initially realized the refinement of quantitative site selection. The results show that : First, the constructed decision tree model can automatically mine the non-linear correlation between multi-dimensional factors and the suitability of old-age care facilities, reduce subjectivity, and take into account scientific rigor and spatial regularity. Second, the accuracy of the simulation results compared with the existing pension facilities is 94.08 %, and the verification method is reliable. Third, the new old-age care facilities in Beijing should focus on the areas with high aging rate within the Third Ring Road, and the remote streets with serious aging outside the Third Ring Road need to increase construction investment.