Journal of Healthcare Engineering / 2018 / Article / Tab 5

Research Article

An Evolutionary Computation Approach for Optimizing Multilevel Data to Predict Patient Outcomes

Table 5

Comparison of results generated from flattened and hierarchical approaches across five patient populations.
(a)

Flattened
ACADCOMMBRAZILUAENHAMCS

Overall AUC0.84310.83610.82610.88200.8429
Training time (hr)42.4778.6719.8915.0029.06
Selected complaints (%)48.352.853.459.049.9

(b)

Hierarchical
ACADCOMMBRAZILUAENHAMCS

Overall AUC0.84330.83640.82600.88190.8436
Training time (hr)4.938.913.463.273.09
Selected complaints (%)49.364.655.655.646.4

(c)

Comparison

Difference in overall AUC ( value)0.61440.22100.70220.36220.2579
Jointly selected complaints (%)28.133.132.437.527.5
Jointly excluded complaints (%)30.627.623.522.931.2