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

Table 2

Comparison of outputs of the MLP ANN algorithms (for the hidden layers 6 × 4 × 2).

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	384.84	1.25	375.04	3.76	389.35	0.09	401.69	3.08
2	379.3	368.24	2.92	367.05	3.23	379.25	0.01	388.01	2.30
3	378.0	365.79	3.23	364.32	3.62	378.13	0.04	369.81	2.17
4	365.0	304.58	16.55	314.20	13.92	364.67	0.09	307.14	15.85
5	377.3	303.08	19.67	312.45	17.19	377.01	0.08	304.51	19.29
6	374.0	303.08	18.96	312.45	16.46	373.65	0.09	304.51	18.58
7	380.3	507.76	33.51	473.41	24.48	380.38	0.02	444.65	16.92
8	368.0	306.38	16.75	316.03	14.12	368.04	0.01	310.58	15.60
9	354.0	315.65	10.83	325.97	7.92	353.46	0.15	321.76	9.11
10	430.0	493.90	14.86	422.77	1.68	429.73	0.06	463.36	7.76
Train error		2.24𝐸−03		2.14𝐸−03		7.77𝐸−04		2.59𝐸−03
Test error		3.17𝐸−03		1.84𝐸−03		6.87𝐸−03		1.85𝐸−03