Table 3. Importance of input variables in LTBI risk prediction model based on SVM algorithm.
| Variant | Importance of forecasting | Variant | Importance of forecasting | Variant | Importance of forecasting |
| Abbreviation: SVM=support vector machines; LTBI=latent tuberculosis infection; BCG=Bacille Calmette–Guérin. | |||||
| Contact type | 0.1476 | Etiological results of index case | 0.1047 | Occupation of index case | 0.0509 |
| Key population classification of index case | 0.1236 | Household registration type | 0.0891 | BCG scar | 0.0281 |
| Residential area of close contact | 0.1202 | Annual household income | 0.0712 | ||
| Frequency of group activity participation | 0.1125 | Weekly frequency of sleep deprivation | 0.0568 | ||