Developing a prediction model for successful aging among the elderly using machine learning algorithms

Area of Science:

• Gerontology
• Artificial Intelligence
• Public Health

Background:

• Global aging trends necessitate strategies for successful aging (SA).
• Quality of Life (QoL) in the elderly is influenced by physical, mental, and social factors.
• Previous research has underemphasized the role of social determinants in SA.

Purpose of the Study:

• To develop a predictive model for successful aging (SA) in the elderly.
• To identify key physical, mental, and social factors contributing to SA.
• To enhance the quality of life (QoL) for the aging population through predictive modeling.

Main Methods:

• Investigated 975 cases of elderly individuals (SA and non-SA).
• Employed univariate analysis to identify significant factors influencing SA.
• Compared performance of machine learning algorithms including Random Forest (RF), XGBoost, J-48, Artificial Neural Network, Support Vector Machine, and Naive Bayes.

Main Results:

• The Random Forest (RF) model achieved the highest predictive performance.
• RF model metrics: PPV=90.96%, NPV=99.21%, sensitivity=97.48%, specificity=97.14%, accuracy=97.05%, F-score=97.31%, AUC=0.975.
• RF demonstrated optimal accuracy in predicting successful aging (SA).

Conclusions:

• Predictive models can significantly improve elderly QoL and reduce societal economic burdens.
• The Random Forest (RF) model is an effective tool for predicting SA in the elderly.
• Incorporating social factors into SA prediction models is crucial for comprehensive elderly care.