Table 4 The statistical errors of the proposed models in train, test, and total data sets.

From: Modeling crude oil pyrolysis process using advanced white-box and black-box machine learning techniques

Statistical criteria

RMSE

SD

R2

MAPE

MBE

MAE

NSE

CatBoost

 Test

0.61533

0.04181

0.99965

2.18740

− 0.00342

0.43954

0.99965

 Train

0.50871

0.03915

0.99976

2.04320

0.00072

0.38990

0.99976

 All

0.92566

0.05094

0.99919

2.76250

− 0.01995

0.63764

0.99919

GPR

 Test

0.42119

0.03885

0.99984

1.88670

0.00696

0.28109

0.99984

 Train

0.42834

0.03992

0.99983

1.93660

0.00529

0.28755

0.99983

 All

0.39137

0.03420

0.99986

1.68790

0.01364

0.25534

0.99986

GP

 Test

4.71424

0.15556

0.97950

10.38880

0.77854

3.37434

0.97915

 Train

4.77625

0.15771

0.97885

10.55290

0.75432

3.42272

0.97849

 All

4.45818

0.14665

0.98204

9.73420

0.87518

3.18126

0.98172

XGBoost

 Test

0.19878

0.01339

0.99996

0.70210

0.00195

0.14107

0.99996

 Train

0.19382

0.01278

0.99997

0.68270

0.00214

0.13864

0.99997

 All

0.21744

0.01559

0.99996

0.77960

0.00120

0.15073

0.99996

Back to article page