Walking behavior patterns of the elderly and its relationship with the built environment considering the activity purpose

Abstract

The proportion of the elderly in the city is rapidly increasing due to the extension of life expectancy and the rapid decrease in the fertility rate. The urban field also focuses on research about the urban environment suitable for elderly walking, which is one of the studies to maximize the health and life of the elderly. In this study, the elderly walking was classified into walking for leisure and exercise (LE), walking for commercial and service (CO), and walking for social (SO) by applying the association rule mining with GPS data of the elderly living in Uijeongbu, Gyeonggi-do, South Korea. The difference in walking behavior was verified according to the classified walking pattern. In addition, the random forest method is used to examine the relationship between walking and the built environment selected by the walking group. It shows that all built environment attributes and path choices have a nonlinear and complex relationship. For example, the ratio of greenness and nature around the path has a positive effect on path selection for exercise and recreation purposes, but generally negative effects on walking for commercial purposes. Even in recreational walking, there is a positive relationship with path selection, especially when the ratio of greenness is more than 15%. Furthermore, an in-depth analysis was conducted by applying the PDP and SHAP methods for interpretation from a global and local perspective. This study is the first to objectively classify the elderly's walking groups according to their activity purpose, prove their differences, and closely analyze the nonlinear relationship with the built environment. It implies that this study will be a valuable guide to efficient space and facility placement, path planning, and design for the elderly in the future.