Table 6 Quantitative performance indicators of models using individual data.

From: Compressive strength prediction of carbonated recycled aggregate concrete using regression based machine learning models

Final test results (Individual data)

Model

Training

Testing

MAE

RMSE

R2

95th percentile error

99th percentile error

Max_error

MAE

RMSE

R2

95th percentile error

99th percentile error

Max_error

Decision Tree

0.7276

0.9032

0.9975

1.7733

2.1495

2.3283

1.1732

1.5914

0.9913

3.4574

3.7499

3.8150

LightGBM

0.7276

0.9032

0.9975

1.7733

2.1495

2.3283

1.1733

1.5914

0.9913

3.4578

3.7503

3.8154

Polynomial MLR (deg = 2)

0.7439

0.9505

0.9972

1.9631

2.1576

2.3507

1.2056

1.6245

0.9910

3.4574

3.7499

3.8150

Random Forest

0.7364

0.9117

0.9975

1.7680

2.0993

2.3577

1.2497

1.6322

0.9909

3.5530

3.8513

3.9164

Ridge

2.6039

3.1980

0.9688

5.8326

6.3379

6.5610

2.2520

2.6890

0.9752

4.8088

4.9973

5.0464

MLR

2.6039

3.1980

0.9688

5.8326

6.3379

6.5611

2.2520

2.6890

0.9752

4.8087

4.9973

5.0465