Related Experiment Video
Updated: Mar 9, 2026
12:18
A Machine Learning Approach to Design an Efficient Selective Screening of Mild Cognitive Impairment
Published on: January 11, 2020
8.2K
Predicting the Livability of Older Adults Using a Livability Scale Through Machine Learning: A Pilot Study
1Department of Occupational Therapy, College of Human Technology Convergence, Woosuk University, Wanju, Republic of Korea.
Evaluation & the Health Professions
|March 7, 2026
Summary
Machine learning accurately predicts residential satisfaction in older adults using the Livability Scale. Random forest models identified key factors like cultural facilities and affordability for better living environments.
Area of Science:
- Gerontology
- Environmental Psychology
- Data Science
Background:
- Understanding livability for older adults is complex, involving person, environment, task, and performance factors.
- Advanced analytical methods are needed to interpret these multidimensional interactions.
- Residential satisfaction is a key indicator of livability in aging populations.
Purpose of the Study:
- To evaluate the predictive accuracy of the Livability Scale (LS) for residential satisfaction in adults aged 65 and above.
- To compare the performance of six machine learning algorithms in modeling livability.
- To identify critical factors influencing residential satisfaction among older adults.
Main Methods:
- Employed six machine learning algorithms: logistic regression, decision tree, random forest, gradient boosting, support vector machine, and an ensemble model.
- Assessed model performance using metrics like F1 score, Area Under the Curve (AUC), sensitivity, and specificity.
- Utilized k-fold cross-validation for robust generalization assessment.
- Conducted feature-importance analysis to pinpoint key predictors of residential satisfaction.
Main Results:
- The random forest model achieved the highest performance (F1 = 0.74, AUC = 0.78), demonstrating strong generalization.
- Feature importance analysis highlighted cultural facilities, affordability, activities, parks, structural safety, and cleanliness as crucial predictors.
- Support vector machine (SVM) showed high sensitivity but more false positives, while gradient boosting machine (GBM) prioritized specificity over sensitivity.
Conclusions:
- Machine learning effectively uncovers non-linear relationships within complex livability data.
- Algorithmic diversity offers valuable insights for developing targeted residential interventions for older adults.
- Key environmental and social factors significantly contribute to predicting older adults' residential satisfaction.
