Skip to main content
. 2021 Jul 22;10(15):3238. doi: 10.3390/jcm10153238

Figure 3.

Figure 3

Data usage. After the data preprocessing, 85% of the data are used for training and 15% are kept for an independent holdout validation of the trained model. KC, keratoconus; 95 CI, 95% confidence interval.