Table 1.

Characterization of neuronal network performance for the prediction of patient outcome. Performance displayed in % for training, testing, and validation samples. The number of training cycle for custom network training is displayed in training algorithm column. BFGS, Broyden–Fletcher–Goldfarb–Shanno training algorithm; SOS, sum of squares.

Prediction target	Net. name	Performance			Training algorithm	Error function	Activation
		Training	Testing	Validation				Train
Infection severity	MLP 89-19-2	100.00	85.71	85.71	BFGS 24	Infection severity	MLP 89-19-2	100.00
Time-to-heal	MLP 89-13-3	91.43	85.71	71.43	BFGS 17	Time-to-heal	MLP 89-13-3	91.43