Hyperparameter Tuning of Machine Learning Algorithms Using Response Surface Methodology: A Case Study of ANN, SVM, and DBN

Table 8

ANOVA of the final model of the second data fold.

Source	Df	Adj SS	Adj MS	F-value	value
Model	4	0.12862	0.032155	10.62	<0.001
HN	1	0.01148	0.011481	3.79	0.070
TC	1	0.08953	0.089534	29.56	<0.001
HN × HN	1	0.01762	0.017617	5.82	0.029
TC × TC	1	0.01483	0.014835	4.9	0.043
Error	15	0.04543	0.003028
Lack-of-fit	10	0.03347	0.003347	1.4	0.373
Pure error	5	0.01195	0.002391
Total	19	0.17405