Predictors of the rate of cognitive decline in older adults using machine learning.

Background: The longitudinal rates of cognitive decline among aging populations are heterogeneous. Few studies have investigated the possibility of implementing prognostic models to predict cognitive changes with the combination of categorical and continuous data from multiple domains.

Objective: Implement a multivariate robust model to predict longitudinal cognitive changes over 12 years among older adults and to identify the most significant predictors of cognitive changes using machine learning techniques.