Enhancing SVM’s robustness of weekly streamflow prediction based on three feature selection algorithms
Bibliographic record
Abstract
Accurate prediction of streamflow is an essential factor for the optimization of water resources management. This study investigates the effectiveness of three feature determination techniques, namely, principal component analysis (PCA), kernel principal component analysis (KPCA) and sequential forward selection (SFS), on Support Vector Machine (SVM) model’ efficiency, for real-world one-week-ahead streamflow forecasting. The study was conducted in three subbasins of Oum-Errbia watershed in Morocco, namely, Ait-Ouchen, Tassaout and Tillouguite for the period going from 2000 to 2019. Initially, an ensemble of 8 input variables (including weekly precipitation (P), difference of P (Pdiff), snow cover area (SCA), difference of SCA (SCAdiff), snow water equivalent (SWE), difference of SWE (SWEdiff), temperature (T), and antecedent streamflow (Q1)), are reduced based on PCA, KPCA and SFS selection. The optimum features are further utilized to optimize the hyper-parameters of SVM model throughout the Grid-Search method. The streamflow simulations obtained by the developed SVM models (PCA-SVM, KPCA-SVM and SFS-SVM) were compared against observed streamflow using performance indicators of Kling-Gupta Efficiency (KGE), Nash-Sutcliffe Efficiency coefficient (NSE), and Root Mean Square Error (RMSE). The results of this study highlight three important findings. First, PCA and SFS techniques were found to be more effective than KPCA in terms of enhancing SVM’ performance in the three sub-basins. Besides, PCA and SFS had approximately the same effect on SVM, they both enhanced its efficiency. Second, KPCA had a negative effect on SVM predictivity over the three subbasins (Ait-Ouchen: NSE= 0.34, 0.43, KGE=0.29, 0.46, RMSE=78.34, 53.31 (MCM), Tassaout: NSE=0.01, 0.00, KGE=-0.30, -0.38, RMSE=8.98, 15.03 (MCM) and Tillouguite: NSE=0.74, 0.57, KGE=0.69, 0.55, RMSE= 48.21, 68.54(MCM)), during the training and the validation, respectively. Finally, the validation results indicated that the highest performance is achieved by PCA-SVM model in Ait-Ouchen (NSE = 0.69, KGE =0.70, RMSE = 38.9409MCM 3), Tassaout (NSE =0.21, KGE =0.10, RMSE = 13.33Mm3) and Tillouguite (NSE = 0.70, KGE = 0.77, RMSE = 57.09MCM). Thus, it is crucial to properly choose the input variables to develop more accurate models for predicting weekly streamflow.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.003 | 0.007 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".