Developing a random forest classifier for predicting the depression and managing the health of caregivers supporting patients with Alzheimer's Disease

Background: Supporting the caregivers of dementia patients is an important issue in the field of public health. Objective: This study established a model for predicting the depression of dementia caregivers while considering the sociodemographic and health science characteristics of South Koreans. The results of this study provided baseline data for developing and applying a caregiver management App. Methods: This study analyzed 2,592 adults (⩾ 19 years old; 1154 men and 1438 women) who were caregivers (e.g., family and caregivers) of demented elderly (⩾ 60 years old). Results: The results of developed random forest model showed that gender, subjective health status, disease or accidence experience within the past two weeks, the frequency of meeting a relative, economic activity, and monthly mean household income were the major predictors for the depression of caregivers. The prediction accuracy of the model was better than K-NN and support vector machine. Conclusions: It was proved that the developed random forest-based App for predicting and managing the depression of dementia caregivers used an algorithm that has a high predictive power. It is required to develop a customized home care system that can prevent and manage the depression of the caregiver.