Table 7 Optimized hyperparameters for support vector regression.

From: Prediction of uniaxial compressive strength of limestone from ball mill grinding characteristics using supervised machine learning techniques

Hyperparameter

Description

Range

Optimal value

Regularization parameter, c

Penalty factor for misclassification error

0.5-8

1

Epsilon, ε

Margin of tolerance where no penalty is given to errors

0.01–0.5

0.01

SVM Kernel

Parameter that influences model’s performance

Linear, polynomial, radial basis function, and sigmoid.

Linear

Back to article page