Skip to main content
. 2023 Dec 22;271(3):1133–1149. doi: 10.1007/s00415-023-12132-z

Fig. 3.

Fig. 3

Performance of random forest algorithm for predicting NEDA and change on therapy during follow-up. The figure shows the ROC curves on the left and the feature importance ranking (top 10) for the best predictor (Table 2) on the right. A staying on NEDA after 2 years follow-up; B starting DMDs during follow-up; C change from the first line to high-efficacy DMDs during follow-up