An approach to construct Bayesian networks for use in living environment field, taking Chinese CCRCs as the example

Motivation: Fast population aging has been considered one of the most challenging demographic transitions globally. The influence of older adults’ living environment on their well-being cannot be ignored.Knowledge gap: Although the importance of living environment factors on the well-being of people has been widely recognized, the interaction between these factors and their interactive influence on the well-being of older people have not been quantified. Aim and objectives: This paper presents a detailed approach to develop a Bayesian network (Bns), focusing on constructing, parameterizing and quantitatively evaluating the Bns, with the ultimate purpose of revealing the impact of living environment on residents' well-being in Chinese Continuing Care Retirement Communities (CCRCs).Research method: Living environment variables that may affect well-being were first identified via comprehensive literature review. Two contextually appropriate scales were then selected to measure well-being. Based on this, we further developed the construction process of Bns for the CCRCs, including qualitative and quantitative parts.Preliminary findings: Fifteen living environment variables were identified, comprising community daily life service, urban infrastructure accessibility, community management and operation, greening space, safety, connection with family, social engagement, community daily support facilities, social connection within the community, walkability, community social activities engagement, community barrier-free infrastructure, health services, and age-friendly indoor environment. The Ryff Scale and the Short Form-36 Scale were considered appropriate for measuring well-being. Combining these findings, the process of developing a definite whole Bns approach for the CCRCs is proposed. Experts' viewpoints and Dempster-Shafer Theory (DST) are utilized to construct Bns. Knowledge Engineering Bayesian Networks (KBEN) are adopted to combine expert and data sources to parameterize and evaluate the Bns.Research significance: This research promotes the innovative application of Bns in environmental research, involving the elements and methods of constructing the model. Researchers can refer to the guidance and use it as a feasible instrument to apply to their research, which advances future investigations in the field