Table 1.
Numbers of divided sample data sets. Only train data for critical thyroid condition are augmented and the original amount of data before the augmentation is given in brackets (N=122,581).
Thyroid conditions | Total number of prepared data sets | ||||
|
Train (n=87,524), n (%) | Validation (n=21,119), n (%) | Test (n=42,237), n (%) | Total, n (%) |
|
Healthy | 29,175 (33.33) | 18,312 (86.71) | 36,624 (86.71) | 84,111 (68.62) |
|
Caution required | 29,175 (33.33) | 2682 (12.70) | 5363 (12.70) | 37,220 (30.36) |
|
Critical | 29,174 [875] (33.33) | 125 (0.59) | 250 (0.59) | 1250 (1.02) |
|