Streamflow forecasting at Segamat Station using a hybrid group method of data handling with particle swarm optimization
Flood forecasting at monsoon-driven stations like the Segamat river station requires accurate streamflow predictions, where conventional Group Method of Data Handling (GMDH) models often underperform due to local optima convergence. This study enhances GMDH through Particle Swarm Optimization (PSO),...
Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Conference or Workshop Item |
| Language: | en |
| Published: |
IEEE
2025
|
| Subjects: | |
| Online Access: | https://umpir.ump.edu.my/id/eprint/47308/1/Streamflow%20Forecasting%20at%20Segamat%20Station%20Using%20a%20Hybrid%20Group%20Method%20of%20Data%20Handling%20with%20Particle%20Swarm%20Optimization.pdf https://umpir.ump.edu.my/id/eprint/47308/ https://doi.org/10.1109/AiDAS67696.2025.11213668 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Flood forecasting at monsoon-driven stations like the Segamat river station requires accurate streamflow predictions, where conventional Group Method of Data Handling (GMDH) models often underperform due to local optima convergence. This study enhances GMDH through Particle Swarm Optimization (PSO), demonstrating significant performance gains. The hybrid GMDH-PSO model achieves a 14.8% average improvement in Nash-Sutcliffe Efficiency (NSE) and 32.5% reduction in Root Mean Square Error (RMSE) compared to standalone GMDH when tested on 63 years of Segamat River data (1960-2022). The optimization reduces Mean Absolute Error by 34.6%, with peak NSE values reaching 0.97 during validation. These quantified improvements confirm PSO's effectiveness in overcoming GMDH's limitations, particularly for extreme flow events where prediction accuracy matters most. The results establish GMDH-PSO as a superior alternative for operational flood forecasting in tropical catchments, with direct implications for early warning systems in vulnerable regions like Johor. |
|---|