Table 3.
Main Performance Results with Interferant Set (Hidden Size = 5000).
| SYNTAX > = 23 | |||
|---|---|---|---|
| Group1 | Group2 | Group3 | |
| AUC-median–%(95% CI) | 82.129 (70.312–91.216) | 89.16 (76.953–98.438) | 78.893 (67.128–89.451) |
| Sensitivity – %(95%CI) | 100.0 (75.0–100.0) | 100.0 (100.0–100.0) | 100.0 (85.294–100.0) |
| Specificity – %(95%CI) | 68.75 (46.875–90.625) | 87.5 (75.0–96.875) | 64.706 (47.059–82.353) |
| Positive predictive value – %(95%CI) | 74.419 (65.306–88.889) | 88.889 (80.0–96.97) | 73.913 (65.385–82.937) |
| Negative predictive value – %(95%CI) | 100.0 (75.0–100.0) | 100.0 (100.0–100.0) | 100.0 (83.321–100.0) |
| Group4 | Group5 | ||
| AUC-median–%(95% CI) | 70.222 (55.331–82.222) | 80.017 (67.904–90.052) | |
| Sensitivity – %(95%CI) | 56.667 (40.0–100.0) | 76.471 (58.824–91.176) | |
| Specificity – %(95%CI) | 93.333 (33.333–100.0) | 88.235 (73.529–97.059) | |
| Positive predictive value – %(95%CI) | 89.474 (60.0–100.0) | 86.207 (74.194–96.552) | |
| Negative predictive value – %(95%CI) | 68.182 (60.417–100.0) | 78.049 (68.182–90.909) | |
| 5 Groups AUC Mean | 80.084 (72.926–85.869) | ||
Bootstrapping is used to estimate 95% confidence intervals (CI) for the performance metrics of our classification results (i.e., AUC, sensitivity, specificity). We apply n-out of-n bootstrap with replacement from our dataset. For each bootstrap sample, we calculate and reserved the performance metrics for that bootstrap sample. The bootstrap sampling is repeated for 2000 times. We then estimate the 95% CI by using the 2.5 and 97.5 percentiles of the empirical distribution of corresponding metrics. All the results retain 3 decimal places, unless the last decimal is 0.
AUC = Area Under the Curve. SYNTAX = Synergy between Percutaneous Coronary Intervention (PCI) with Taxus and Cardiac Surgery.