Application of Machine Learning Methods in Nursing Home Research.

Area of Science:

• Gerontology
• Computer Science
• Healthcare Informatics

Background:

• Machine learning (ML) systems enhance prediction and knowledge generation through data-driven insights.
• Predictive modeling is crucial for improving healthcare outcomes in long-term care facilities.

Purpose of the Study:

• To compare the predictive accuracy of six machine learning methods for falls in nursing homes (NHs).
• To identify the most effective ML algorithm for fall risk assessment in NH settings.

Main Methods:

• Applied six ML algorithms: random forest (RF), logistic regression, and three Support Vector Machine (SVM) variants (linear, polynomial, radial, sigmoid).
• Utilized a preprocessed dataset of 60 individuals for model development.
• Evaluated model performance using an accuracy metric.

Main Results:

• Random Forest (RF) achieved the highest accuracy (0.883) in predicting falls.
• Logistic regression, linear SVM, and polynomial SVM showed comparable accuracy (0.867).

Conclusions:

• Random Forest (RF) is a powerful tool for identifying fall predictors in nursing home environments.
• Effective fall management requires considering both organizational and personal factors.
• Further research should explore additional factors and advanced ML methods to confirm ML's utility in NH research.