Table 10 Comparison of proposed models with literature.

From: Development of machine learning models for forecasting the strength of resilient modulus of subgrade soil: genetic and artificial neural network approaches

Models

Phases

R2

RMSE (MPa)

MAE (MPa)

PI

Reference

ANN

Training

0.97

15.97

11.32

Indraratna et al.110

Testing

0.93

25.12

18.17

Validation

0.94

22.78

17.11

ANFIS

Training

0.97

15.11

10.41

Testing

0.87

32.70

21.29

Validation

0.92

27.39

18.40

MLR

Training

0.32

0.10

0.07

0.54

Kardani et al.28

GBR

Training

0.77

0.09

0.06

1.18

DTR

Training

0.91

0.04

0.03

1.78

KNR

Training

0.54

0.09

0.04

1.00

RFR

Training

0.93

0.03

0.02

1.83

VO-ENSM

Training

0.88

0.06

0.04

1.61

VO-ENSM (RF)

Training

0.92

0.05

0.03

1.72

ST-ENSM

Training

0.92

0.04

0.02

1.80

BG-ENSM

Training

0.98

0.02

0.01

1.94

MLR

Testing

0.29

0.10

0.07

0.47

GBR

Testing

0.79

0.09

0.06

1.19

DTR

Testing

0.93

0.03

0.02

1.82

KNR

Testing

0.73

0.07

0.03

1.38

RFR

Testing

0.95

0.03

0.02

1.87

VO-ENSM

Testing

0.92

0.05

0.03

1.70

VO-ENSM (RF)

Testing

0.95

0.04

0.03

1.78

ST-ENSM

Testing

0.95

0.03

0.02

1.87

BG-ENSM

Testing

1.00

0.01

0.00

1.99

MLR

Testing

0.87

2143.00

Pahno et al.111

Regression Tree

Testing

0.84

2339.00

RF

Testing

0.94

1499.00

XGBoost

Testing

0.95

1321.00

GEP

Training

0.99

2.41

1.72

0.04

This Study

Testing

0.98

3.64

2.57

0.05

MEP

Training

0.98

3.82

2.50

0.06

Testing

0.97

4.89

3.28

0.07

ANN

Training

0.97

4.94

3.10

0.07

Testing

0.95

7.38

5.48

0.11