Skip to main content
. Author manuscript; available in PMC: 2021 Jan 15.
Published in final edited form as: Neuroimage. 2020 Oct 16;225:117451. doi: 10.1016/j.neuroimage.2020.117451

Fig. 8. Most important variables from the All Sequences (TS) 50 mm3 cutoff for the 1000 training and validation set reassignments.

Fig. 8.

The plots show smoothed data from the three most important RIA features for the All Sequence (TS) modeling scenario over the 1000 reassignments to the training and validation set results from Table 4. The top row is for the first-step classification model and the bottom row the second-step regression model. Smoothing was performed with a loess smoother.