Table 6 Results of multi-step ahead EC forecasting for the Albert River.

From: Multi-step ahead forecasting of electrical conductivity in rivers by using a hybrid Convolutional Neural Network-Long Short-Term Memory (CNN-LSTM) model enhanced by Boruta-XGBoost feature selection algorithm

Horizon

Data

R

RMSE

MAPE

E

Tstat

U95%

3-steps-ahead

Train

0.8947

73.6800

10.4113

0.7998

2.3851

204.1362

Test

0.8764

66.3651

12.0275

0.7633

4.7504

183.0642

5-steps-ahead

Train

0.8674

81.9761

12.8438

0.7521

1.8250

227.1739

Test

0.8326

75.6831

14.0573

0.6922

1.4061

209.7361

7-steps-ahead

Train

0.8404

90.1850

15.0560

0.6994

6.1947

249.0694

Test

0.7934

83.1531

16.8054

0.6284

0.1597

230.5399

10-steps-ahead

Train

0.8004

98.5680

17.4178

0.6402

0.2013

273.2424

Test

0.7367

92.7799

20.2669

0.5374

3.5644

256.4918

Back to article page