Fig. 3
From: Predicting self-perceived general health status using machine learning: an external exposome study
Relationship between variable importance (VI) score of variables present in both the 2012 and 2016 RF model. Legend: The VI score expressed as the difference between the original error of the model and the error after randomization of the variable of interest. The VI procedure was performed using the iml package in R in triplicate, and the average score is presented. Point color represents the type of variable: personal characteristics, neighborhood characteristics and environmental characteristic