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

Table 12

Best hyperparameter settings for dot and radial kernel functions using RSM.

Fold	Dot			Radial
	C	Convergence	MAE (validation)	C	Convergence	MAE (validation)
1	12	0.001	1.1680	12	0.0109	1.1455
2	−1	0.1	1.1868	3	0.001	1.1237
3	−1	0.0406	1.1102	−1	0.0703	1.1575
4	-1	0.001	1.1694	6	0.001	1.0529
5	11	0.0109	1.1209	9	0.001	1.1137
6	9	0.0307	1.0528	17	0.0208	1.0892
7	−1	0.1	1.2577	11	0.001	1.2452
8	−1	0.0505	1.1234	3	0.0109	1.1604
9	−1	0.0901	1.0765	9	0.001	1.0988
10	10	0.0406	1.1923	7	0.1	1.1827