• The Journal of the Convergence on Culture Technology
• /
• v.10 no.6
• /
• pp.677-689
• /
• 2024

The quality of life, frailty, economic activity, and other indicators are crucial for assessing older adults' overall well-being and health status. A comprehensive evaluation using this information helps predict the health status of older adults. This study aims to apply and compare machine learning-based prediction models for comprehensive health indicators of community-dwelling older adults. Utilizing data from 4,652 individuals provided by the Aging Research Panel, we assessed various machine learning techniques to fit the predictor variables. Our findings reveal that the LightGBM Regression model performed the best, with an RMSE of 5.082 and an MSE of 25.83. The Gradient Boosting model best predicted current health status, with an RMSE of 0.588 and an R-Square of 0.456. Additionally, the Random Forest model showed strong performance in predicting economic activity participation among older adults. These machine learning-based models offer valuable insights for evaluating health status and predicting economic activity participation, highlighting the importance of employing diverse methodologies for comprehensive predictions.