Table 3 False classifications (%) with jackknife cross-validation by optimized Classification And Regression Tree (CART) models
From: Optimizing the diagnostic power with gastric emptying scintigraphy at multiple time points
Type of Misdiagnosis | 1-h | 2-h | 3-h | 4-h | 2-h + 4-h | 3-h + 4-h | 4-h + 3-h + 2-h + 1-h + Age |
|---|---|---|---|---|---|---|---|
| 76 (61.8%) | 28 (22.8%) | 29 (23.6%) | 28 (22.8%) | 28 (22.8%) | 28 (22.8%) | 26 (21.1%) |
| 24 (12.2%) | 69 (35.0%) | 44 (22.3%) | 18 (9.1%) | 18 (9.1%) | 18 (9.1%) | 18 (9.1%) |
Total Misdiagnosis | 100 (31.3%) | 97 (30.3%) | 73 (22.8%) | 46 (14.4%) | 46 (14.4%) | 46 (14.4%) | 44 (13.8%) |
AUC For Optimized Model | 0.724 | 0.753 | 0.825 | 0.867 | 0.865 | 0.858 | 0.881 |
- Note:
: model predicted disease status (1 for yes, 0 for no); - y: diagnosed disease status (1 for yes, 0 for no).
: model predicted disease status (1 for yes, 0 for no);