Low-Code Application and Practical Implications of Common Machine Learning Models for Predicting Punching Shear Strength of Concrete Reinforced Slabs

Table 3

Hyperparameters with tuning range and optimal values.

Model	Hyperparameter name	Range	Optimal value using Bayesian process	Optimal value using random search
GPR	“Sigma”	[1.00e−04,3.78e + 03]	41.401	38.7
	“BasisFunction”	“constant” “none”	“linear”	“none”
		“linear”
		“pureQuadratic”
	“KernelFunction”	“ardmatern32”	“matern32”	“matern32”
		“ardmatern52”
		“exponential”
		“matern32” “matern52”
		“squaredexponential”
	“KernelScale”	(1.6870, 1,687)	7.329	18.3
	“Standardise”	“true” “false”	“true”	“true”
SVM	“BoxConstraint”	(1.00e−03, 1,000)	83.518	68.53
	“Epsilon”	(0.165, 1.65e + 04)	0.293	0.282
	“KernelFunction”	“gaussian” “linear” “polynomial”	“polynomial”	“polynomial”
	“PolynomialOrder”	[2, 4]	2	2
	“Standardise”	“true” “false”	“true”	“true”
Ensemble	“Method”	“Bag “LSBoost”	“LSBoost”	“LSBoost”
	“NumLearningCycles”	(10,500)	487	497
	“LearnRate”	(1.00e−03,1)	0.049	0.035
	“MinLeafSize”	(1,152)	2	4
	“MaxNumSplits”	(1,303)	5	103
	“NumVariablesToSample”	[1, 6]	5	4