The Estimation of Product Standard Time by Artificial Neural Networks in the Molding Industry

Table 3

Comparison of outputs of the MLP ANN algorithms (for the hidden layers 7 × 5 × 3).

No.	Pre	Generated	Absolute	Generated	Absolute	Generated	Absolute	Generated	Absolute
	measur-	values by	difference	values by	difference	values by	difference	values by	difference
	ed values	CGF	(%)	BFG	(%)	LM	(%)	BR	(%)
1	389.7	385.75	1.01	446.06	14.46	350.30	10.11	402.12	3.19
2	379.3	368.12	2.95	415.30	9.49	385.63	1.67	388.42	2.40
3	378.0	399.98	5.82	436.44	15.46	449.68	18.96	370.51	1.98
4	365.0	315.42	13.58	278.53	23.69	187.77	48.56	306.98	15.90
5	377.3	307.55	18.49	273.95	27.39	115.89	69.29	304.36	19.33
6	374.0	307.54	17.77	273.95	26.75	115.89	69.01	304.36	18.62
7	380.3	365.49	3.89	615.83	61.93	407.08	7.04	444.74	16.94
8	368.0	326.87	11.18	285.57	22.40	278.50	24.32	310.39	15.65
9	354.0	353.95	0.01	314.61	11.13	395.49	11.72	321.62	9.15
10	430.0	477.81	11.12	453.62	5.49	420.70	2.16	462.88	7.65
Train error		1.85𝐸−03		1.48𝐸−03		3.00𝐸−24		2.59𝐸−03
Test error		1.31𝐸−03		7.96𝐸−03		1.46𝐸−02		1.85𝐸−03