Table 4. Model performances for models trained on all features restricted to tp0, tp1, tp2, tp0 + tp1 and all timepoints using 5-fold cross-validation.
Timepoints | AUC | Precision | Recall |
---|---|---|---|
Tp0 | 0.918 (0.856, 0.980; p<0.001) | 0.778 (0.716, 0.840; p<0.001) | 0.913 (0.885, 0.941; p<0.001) |
Tp1 | 0.844 (0.705, 0.983; p = 0.004) | 0.793 (0.653, 0.933; p = 0.007) | 0.841 (0.709, 0.973; p = 0.003) |
Tp2 | 0.920 (0.858, 0.982; p<0.001) | 0.832 (0.77, 0.894; p<0.001) | 0.951 (0.929, 0.973; p<0.001) |
Tp0 + Tp1 | 0.938 (0.89, 0.986; p<0.001) | 0.805 (0.757, 0.853; p<0.001) | 0.963 (0.918, 1.008; p<0.001) |
Tp0 + Tp1 + Tp2 | 0.951 (0.910, 0.992; p<0.001) | 0.815 (0.690, 0.940; p = 0.004) | 0.926 (0.844, 1.008; p<0.001) |
Values in parentheses are 95% confidence intervals. P-values are probability of AUC being 0.5. AUC: area under the curve; Tp: time point.