Table 4 Discrimination and Calibration Performance in the Training and Testing Sets for Multivariate Adaptive Regression Spline Algorithm
Training set (cross-validation) | Testing set | |
|---|---|---|
Discrimination performance | ||
AUROC | 0.84 (95% CI 0.82–0.86) | 0.88 (95% CI 0.83–0.93) |
Accuracy | 77.6% (95% CI 75.6%—79.5%) | 81.2% (95% 74.9%—86.4%) |
No-information rate | 53.7% | 53.9% |
p-value [accuracy > no-information rate]* | < 0.001 | < 0.001 |
p-Value [Chi-Square goodness of fit] | 0.10 | 0.08 |
Calibration performance | ||
Spiegelhatler’s Z-score | 0.83 | -1.31 |
p-value | 0.20 | 0.09 |