Table 4. Results for different workflows of logistic regression on hospital readmissions data, with standard deviations over n = 10 runs.

LR Workflow	ROC AUC	Sensitivity	Specificity	Accuracy
No Transformation	0.49 (0.023)	0.58 (0.031)	0.48 (0.021)	0.49 (0.022)
No Transformation, SMOTE	0.64 (0.033)	0.62 (0.029)	0.67 (0.039)	0.66 (0.037)
No Transformation, ROSE	0.53 (0.041)	0.54 (0.044)	0.51 (0.045)	0.52 (0.045
PCA	0.58 (0.017)	0.68 (0.022)	0.45 (0.029)	0.49 (0.028)
PCA, SMOTE	0.49 (0.037)	0.64(0.035)	0.44 (0.034)	0.47 (0.034)
PCA, ROSE	0.45 (0.061)	0.50 (0.059)	0.55(0.065)	0.54 (0.063)
Mapper, No Transformations	0.61 (0.048)	0.62 (0.052)	0.53 (0.049)	0.55 (0.050)
Mapper, No Transformations, SMOTE	0.67 (0.066)	0.60 (0.055)	0.60 (0.064)	0.60 (0.062)
Mapper, No Transformation, ROSE	0.62 (0.073)	0.69 (0.076)	0.59 (0.078)	0.61 (0.078)
Mapper, Node PCA	0.55 (0.065)	0.62 (0.058)	0.50 (0.059)	0.52 (0.058)
Mapper, Node PCA, SMOTE	0.69(0.071)	0.62 (0.069)	0.78 (0.065)	0.75 (0.066)
Mapper, Node PCA, ROSE	0.61 (0.084)	0.58 (0.082)	0.63 (0.087)	0.62 (0.086)