Table 8 Performance evaluation of random Forest, XG Boost, KNN, and ANN models during the training and testing phases.

From: Experimental and machine learning prediction of compressive strength of chemically activated RHA based RAC using SHAP and PDP analysis

 

Evaluation parameters

Random forest

XG Boost

KNN

ANN

Training phase (80%)

R2

0.994

0.998

0.735

0.983

MSE

2.132

0.515

89.28

5.654

RMSE

1.460

0.718

9.449

2.378

NRMSE

0.017

0.008

0.109

0.027

MAE

0.840

0.54

7.201

1.824

MAPE %

2.211

1.489

24.55

6.142

Testing phase (20%)

R2

0.925

0.951

0.559

0.923

MSE

15.87

10.38

93.81

16.46

RMSE

3.984

3.222

9.686

4.058

NRMSE

0.075

0.061

0.183

0.077

MAE

2.310

1.862

7.046

2.872

MAPE %

5.815

4.299

21.19

7.184

Back to article page