Evaluation of tree regression analysis for estimation of river basin discharge
| Autor: | Ahmed Mohammed Sami Al-Janabi, Aminuddin Ab Ghani, Parveen Sihag, Nashwan K. Alomari, Somvir Singh Nain |
|---|---|
| Rok vydání: | 2021 |
| Předmět: |
geography
geography.geographical_feature_category 010504 meteorology & atmospheric sciences Correlation coefficient Mean squared error Discharge 0207 environmental engineering Drainage basin 02 engineering and technology 01 natural sciences Random forest Water resources Statistics Autoregressive integrated moving average Computers in Earth Sciences Statistics Probability and Uncertainty 020701 environmental engineering General Agricultural and Biological Sciences 0105 earth and related environmental sciences General Environmental Science Test data Mathematics |
| Zdroj: | Modeling Earth Systems and Environment. 7:2531-2543 |
| ISSN: | 2363-6211 2363-6203 |
| Popis: | River discharge links the hydrologic and geologic cycles in addition to climate components; therefore, it forms an important source of hydraulic and hydrologic quantity. The ability to quantify river discharge accurately is very important for estimating water availability and distribution for better water resources management. In this study, the performance of ARIMA, random forest (RF), the M5P and Bagged M5P (BM5P) methods, for modeling the daily discharge of the Baitarani Riverwere compared and evaluated against measured values. Fifteen different input combinations under two groups (i.e., discharge and rainfall) were considered, and a suitable modeling approach with appropriate model input combination is proposed on the basis of various goodness fit parameters. Four statistical assessment methods implemented to determine the best performing models include the correlation coefficient (CC), Mean square error (MSE), Root mean square error (RMSE) and Scattering Index (SI).The outcomes of this study indicated that the Bagged M5P modeling approach is outperforming than ARIMA, RF and M5P. This model recorded up to 0.8676, 10.7279, 39.836 m3/s and 0.9599 for (CC), (MAE), (RMSE) and (SI), respectively, for testing data set. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|
Full text from SpringerLink