[Retracted] The Comprehensive Evaluation Model of Business Performance Based on Artificial Neural Network

Table 2

Comparison of the errors of the three training algorithms.

Training algorithm	Data set	Preset number of iterations	Mean square error	Mean error
LM	Training set	1000	5.99E − 07	0.0037
	Test set		1.58E − 04	0.071
Momentum gradient descent	Training set	1000	6.07E − 04	0.0088
	Test set		2.08E − 04	0.070
Bayesian regularization	Training set	1000	9.04E − 15	0.0015
	Test set		6.41E − 04	0.022