Table 3 Tuned hyperparameters of the machine learning (ML) models.

From: A novel approach of developing machine learning based models for the prediction of facial dimensions from dental parameters

ML Models

Targets

Best hyperparameters

Support Vector Regression (SVR) Model

Al-Al

‘C’: 0.1, ‘degree’: 2, ‘epsilon’: 0.1, ‘kernel’: ‘rbf’

Ch-Ch

‘C’: 0.1, ‘degree’: 2, ‘epsilon’: 0.1, ‘kernel’: ‘rbf’

Ft-Ft

‘C’: 0.1, ‘degree’: 2, ‘epsilon’: 0.5, ‘kernel’: ‘rbf’

Go-Go

‘C’: 0.1, ‘degree’: 2, ‘epsilon’: 0.1, ‘kernel’: ‘rbf’

Ic -Ic

‘C’: 0.1, ‘degree’: 2, ‘epsilon’: 0.1, ‘kernel’: ‘rbf’

Oc-Oc

‘C’: 0.1, ‘degree’: 2, ‘epsilon’: 0.5, ‘kernel’: ‘linear’

Pu-Pu

‘C’: 0.1, ‘degree’: 3, ‘epsilon’: 0.1, ‘kernel’: ‘poly’

Zy-zy

‘C’: 1, ‘degree’: 2, ‘epsilon’: 0.5, ‘kernel’: ‘linear’

Random Forest

Regression (RFR) Model

Al-Al

‘Max_depth’: 10,‘min_samples_split’: 5,‘n_estimators’: 200

Ch-Ch

‘max_depth’: None,‘min_samples_split’: 5,‘n_estimators’: 50

Ft-Ft

‘max_depth’: 10, ‘min_samples_split’: 5, ‘n_estimators’: 200

Go-Go

‘max_depth’: 10, ‘min_samples_split’: 5, ‘n_estimators’: 200

Ic -Ic

‘max_depth’: 10, ‘min_samples_split’: 2, ‘n_estimators’: 200

Oc-Oc

‘max_depth’: 10, ‘min_samples_split’: 2, ‘n_estimators’: 100

Pu-Pu

‘max_depth’: 10, ‘min_samples_split’: 5, ‘n_estimators’: 200

Zy-zy

‘max_depth’: 10, ‘min_samples_split’: 5, ‘n_estimators’: 50

Decision Tree Regression (DTR) Model

Al-Al

‘max_depth’: 10, ‘min_samples_split’: 2

Ch-Ch

‘max_depth’: 10, ‘min_samples_split’: 10

Ft-Ft

‘max_depth’: 10, ‘min_samples_split’: 5

Go-Go

‘max_depth’: 10, ‘min_samples_split’: 10

Ic -Ic

‘max_depth’: 10, ‘min_samples_split’: 10

Oc-Oc

‘max_depth’: 10, ‘min_samples_split’: 10

Pu-Pu

‘max_depth’: 10, ‘min_samples_split’: 10

Zy-zy

‘max_depth’: 10, ‘min_samples_split’: 10

  1. *C = Regularization parameters, rbf = Radial Basis Function. (Linear regression model do not contain hyperparameters.
Back to article page