A Study On The Spatial Accessibility Of Elderly Meal-Aid Facilities And Its Influencing Factors In Shanghai

Currently,the community-based elderly care model is responsible for providing the majority of elderly care services.As an important component of this model,the spatial distribution of meal-aid facilities must match the needs of the elderly population to ensure that they can enjoy convenient and affordable dining services.However,research on elderly meal assistance is limited,especially regarding the matching degree between meal assistance facilities and the elderly population.There is a lack of diverse research methods,weak theoretical research,and insufficient empirical analysis in related studies.The scientific problem addressed in this paper is how to evaluate whether the service capacity of meal assistance facilities in Shanghai matches the actual needs of the elderly population from the perspective of a living circle scale.Furthermore,how to define segmentation standards to evaluate the accessibility level of meal assistance,and how to calculate the current number of meal assistance facility deficiencies in various neighborhoods.Based on these research questions,this study uses a Gaussian two-step floating catchment area method to calculate the accessibility of meal assistance for nine different time thresholds of research units.For the accessibility calculation values,descriptive statistics,accurate delimitation,and facility deficiency calculations are performed.Analysis of aggregate data is conducted based on multiple administrative levels and spatial autocorrelation accessibility numerical spatial aggregation patterns.Multiple dimensions of influencing factors are constructed based on multisource data,and a random forest model is used for modeling,feature importance recognition,and feature marginal effect interpretation.Finally,spatial heterogeneity research is conducted on the six factors using a geographically weighted regression method.The results show that the adaptation of meal assistance services to the elderly population exhibits spatial differentiation characteristics.Most neighborhoods are below the meal assistance accessibility standard line,and the number of meal assistance facilities needed to reach the standard varies between 2 and 14 among neighborhoods.The deficiencies of meal assistance facilities are more severe in the west side of Puxi,especially in the east side area near the Huangpu River.Overall,the northwest region is better than other regions,showing a west-strong and east-weak trend.Neighborhoods with low-low aggregation of accessibility are mainly concentrated in areas inside the outer ring of Shanghai,while neighborhoods with high-high aggregation are mainly concentrated in areas dominated by Jiading District.The random forest model,based on 21 feature factors,has an R-squared value of 0.89 on the test set,demonstrating excellent performance.The accessibility of daytime care centers and land-use mix show a positive linear impact on meal assistance accessibility prediction,while the accessibility of nursing homes,elderly care homes,and average house prices exhibit nonlinear effects.The six influencing factors all show different degrees of spatial heterogeneity characteristics,with the accessibility of daytime care centers and meal assistance exhibiting a one-way impact,and the remaining five factors exhibiting a two-way impact.