Machine learning algorithm and neural network architecture for optimization of pharmaceutical and drug manufacturing industrial effluent treatment using activated carbon derived from breadfruit (Treculia africana)

Journal of Engineering and Applied Science

Table 4 ANOVA of response surface design of the treatment process

Source	Sum of squares	df	Mean square	F-value	p-value
Model	1201.42	8	150.18	44.62	< 0.0001	Significant
A—Contact time	142.96	1	142.96	42.48	< 0.0001
B—Dosage	100.39	1	100.39	29.83	0.0001
C—pH	1.33	1	1.33	0.3946	0.5416
D—Particle size	21.82	1	21.82	6.48	0.0256
AB	11.92	1	11.92	3.54	0.0844
AC	49.29	1	49.29	14.64	0.0024
BC	40.40	1	40.40	12.00	0.0047
A²	53.28	1	53.28	15.83	0.0018
Residual	40.39	12	3.37
Lack of fit	15.89	9	1.77	0.2162	0.9680	Not significant
Pure error	24.50	3	8.17
Cor. total	1241.81	20