Table 2 Shows the performance (mean of ten-fold cross-validation) for each performance parameter with 95% confidence intervals based on the LOS class
Classifiers | Accuracy [95% Cl] | Precision [95% Cl] | Recall [95% Cl] | F1-Score [95% Cl] | AUC [95% Cl] |
|---|---|---|---|---|---|
LR | 0.8952 (0.8948, 0.8956) | 0.8356 (0.8298, 0.8414) | 0.8952 (0.8948, 0.8956) | 0.8477 (0.8474, 0.8481) | 0.8088 (0.8020, 0.8157) |
ANNs | 0.8240 (0.8182, 0.8299) | 0.8718 (0.8695, 0.8741) | 0.8240 (0.8182, 0.8299) | 0.8434 (0.8400, 0.8468) | 0.8135 (0.8082, 0.8188) |
SVM | 0.8216 (0.8154, 0.8277) | 0.8721 (0.8701, 0.8742) | 0.8216 (0.8154, 0.8277) | 0.8419 (0.8376, 0.8463) | 0.8111 (0.8058, 0.8165) |
RF | 0.8863 (0.8837, 0.8889) | 0.8487 (0.8450, 0.8523) | 0.8863 (0.8837, 0.8889) | 0.8601 (0.8578, 0.8623) | 0.8075 (0.8037, 0.8114) |
XGBoost | 0.8841 (0.8831, 0.8852) | 0.8590 (0.8584, 0.8596) | 0.8841 (0.8831, 0.8852) | 0.8631(0.8616, 0.8646) | 0.8143 (0.8094, 0.8191) |