Table 5.

Statistical data from the three artificial neural network models. Performance shown is based on the mean squared error (MSE).

Stage	Samples × Participants	Accuracy	Error	Performance (MSE)
Model 1: Asians
Training	134	97%	3%	0.03
Testing	34	71%	29%	0.24
Overall	168	92%	8%	-
Model 2: Non-Asians
Training	216	99%	1%	0.01
Testing	54	76%	24%	0.20
Overall	270	94%	6%	-
Model 3: General
Training	350	97%	3%	0.03
Testing	88	71%	29%	0.25
Overall	438	89%	11%	-