Skip to main content
. 2020 Sep 30;10(3):100668. doi: 10.1016/j.imr.2020.100668

Fig. 4.

Fig. 4

Prediction performance comparison with a previous method. (a) Performance of the ERT model trained with a dataset of 1338 subjects with 248 features. (b) Performance of the Mlogit model trained with a dataset of 1338 subjects with 15 features. SE, So-Eum type; SY, So-Yang type; TE, Tae-Eum type.