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

Table 1

List of ANN hyperparameters and their ranges.

ML	Hyperparameter	Description	Lower	Upper	No. of steps
ANN	HN	No. of hidden nodes	1	10	10
	TC	No. of training cycles	10	1000	21
	LR	Learning rate	0.0001	0.1	21
SVM	C	Complexity constant	−1	100	102
	Conv	Convergence epsilon	0.001	0.1	11
DBN	HN	No. of hidden nodes	1	10	3
	ni	No. of pretraining cycles	10	1000	3
	n	No. of training cycles	10	1000	3
	LRRBM	Pretraining learning rate	0.0001	0.1	3
	LR	Learning rate	0.0001	0.1	3
	Nb	Batch size	350	1700	3