Table 10 Multi-step-ahead forecasting for Kon Tum dataset.

From: Accurate discharge and water level forecasting using ensemble learning with genetic algorithm and singular spectrum analysis-based denoising

Metrics/models

Ensemble SSA GA

Ensemble SSA

Ensemble

LSTM

ANN

ARIMA

Q

NSE

0.611

0.549

0.508

0.428

0.246

0.396

MSE

4746.415

4919.71

5365.020

8118.964

10,712.340

7251.454

MAE

25.738

32.909

38.824

41.681

38.596

36.815

MAPE

21.048

33.637

48.5814

45.503

37.246

33.725

H

NSE

0.815

0.804

0.792

0.667

0.626

0.685

MSE

0.053

0.056

0.059

0.099

0.110

0.087

MAE

0.131

0.143

0.165

0.196

0.183

0.171

MAPE

0.025

0.027

0.032

0.038

0.035

0.033

  1. Best results are highlighted in bold text.
  2. Our proposed method (i.e., Ensemble SSA GA) outperforms the others.
Back to article page