Table 1 |
||||||||
| The numbers of excluded and used data for the training and test of each individual diagnostic model | ||||||||
| List of exclusion reasons (Total no. of subjects: training 2462; test 511) | Training | Test | ||||||
| Face | Body shape | Voice | Questionnaire | Face | Body shape | Voice | Questionnaire | |
| Initially not collected | 22 | 0 | 409 | 409 | 0 | 0 | 0 | 0 |
| Low quality pictures or voice recording files | 133 | - | 132 | - | 6 | - | 14 | - |
| TY samples | 58 | 59 | 54 | 56 | 6 | 6 | 6 | 6 |
| Subjects younger than 15 years old | 52 | 54 | 39 | 41 | 13 | 13 | 11 | 13 |
| Data extraction errors and missing cases | 1189 | 97 | 3 | 0 | 132 | 0 | 0 | 0 |
| Outliers considered in the calculation of the moving average | 46 | 27 | 86 | 16 | 0 | 7 | 26 | 0 |
| Excluding influential samples | 6 | 11 | 49 | 25 | 0 | 0 | 0 | 0 |
| Final no. of subjects | 956 | 2214 | 1687 | 1915 | 354 | 485 | 480 | 492 |
| Common samples after integrating the four components (for the integrated diagnostic model) | 729 (male: 241, female: 488) | 346 (male: 114, female: 232) | ||||||
Do et al. BMC Complementary and Alternative Medicine 2012 12:85 doi:10.1186/1472-6882-12-85
