Ridge Regression

The psychological bond between elderly residents and their living environment—termed place attachment—plays a critical role in aging-in-place strategies. This study investigates the impact of environmental design characteristics on place attachment and evaluates the predictive capabilities of machine learning in this context. Methods: A cross-sectional survey was conducted among 490 elderly residents in Tehran using a 38-item Likert-scale questionnaire. The study applied three regression-based algorithms—Linear, Polynomial, and Ridge Regression—to model the relationship between 20 environmental design variables and place attachment scores. Results: "Positive Home Experiences" (r = 0.68), "Freedom from Confinement" (r = 0.64), and "Safety Features" (r = 0.53) emerged as the most influential predictors. Ridge Regression achieved the highest prediction accuracy, with an R² value of 0.6792. Conclusion: The findings demonstrate the potential of machine learning to support human-centered design by enabling the early-stage evaluation of housing for the elderly. The proposed predictive framework can inform architecture curricula, computer-aided design (CAD) tools, and age-friendly housing policies.