Enhancing the performance of data-driven models for monthly reservoir evaporation prediction
The accuracy level for reservoir evaporation prediction is an important issue for decision making in the water resources field. The traditional methods for evaporation prediction could encounter numerous obstacles owing to the effect of several parameters on the shape of the evaporation pattern. The...
Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Published: |
2021
|
| Subjects: | |
| Online Access: | http://eprints.um.edu.my/27011/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | The accuracy level for reservoir evaporation prediction is an important issue for decision making in the water resources field. The traditional methods for evaporation prediction could encounter numerous obstacles owing to the effect of several parameters on the shape of the evaporation pattern. The current research presented modern model called the Coactive Neuro-Fuzzy Inference System (CANFIS). Modification for such model has been achieved for enhancing the evaporation prediction accuracy. Genetic algorithm was utilized to select the effective input combination. The efficiency of the proposed model has been compared with popular artificial intelligence models according to several statistical indicators. Two different case studies Aswan High Dam (AHD) and Timah Tasoh Dam (TTD) have been considered to explore the performance of the proposed models. It is concluded that the modified GA-CANFIS model is better than GA-ANFIS, GA-SVR, and GA-RBFNN for evaporation prediction for both case studies. GA-CANFIS attained minimum RMSE (15.22 mm month(-1)for AHD, 8.78 mm month(-1)for TTD), minimum MAE (12.48 mm month(-1)for AHD, 5.11 mm month(-1)for TTD), and maximum determination coefficient (0.98 for AHD, 0.95 for TTD). |
|---|