Fig. 3.

Performance of models with different weights to balance reinforcement learning and supervised learning in 5-fold cross validation. The dark lines in the middle indicate the mean values, and the shaded areas indicate 1 standard deviation above and below the mean values