Belarbi et al. (2025) Efficient Hyperparameter Optimization for Reference Evapotranspiration Estimation with Limited Parameters: A Comparison of Optuna and Grid Search in the Doukkala Region, Morocco

⚠️ Warning: This summary was generated from the abstract only, as the full text was not available.

Identification

• Journal: Springer Link (Chiba Institute of Technology)
• Year: 2025
• Date: 2025-12-19
• Authors: Zaid Belarbi, Yacine El Younoussi
• DOI: 10.1051/e3sconf/202568000076/pdf

Research Groups

• Not explicitly mentioned in the provided text (Study conducted in the Doukkala region, Morocco).

Short Summary

This study evaluates the performance of four machine learning models for daily reference evapotranspiration (ETo) estimation in a semi-arid region using limited meteorological data. The research demonstrates that the Optuna optimization framework provides a more efficient and effective alternative to Grid Search for hyperparameter tuning.

Objective

• To assess and compare the accuracy of machine learning models (SVR, RF, XGB, and ANN) for daily ETo estimation using limited input parameters and to evaluate the efficiency of two hyperparameter optimization strategies (Grid Search vs. Optuna).

Study Configuration

• Spatial Scale: Doukkala region, Morocco (semi-arid climate).
• Temporal Scale: Daily resolution.

Methodology and Data

• Models used: Support Vector Regression (SVR), Random Forest (RF), Extreme Gradient Boosting (XGB), and Artificial Neural Networks (ANN).
• Data sources: Meteorological data including Julian day, maximum temperature ($T{max}$), minimum temperature ($T{min}$), and relative humidity.

Main Results

• All models achieved high predictive accuracy with $R^2$ values between 0.90 and 0.916 and RMSE values ranging from 0.53 to 0.56 mm/day.
• Optuna consistently matched or outperformed Grid Search while requiring fewer evaluations; for instance, XGB performance improved from $R^2 = 0.9040$ (Grid Search) to $R^2 = 0.9137$ (Optuna).
• Random Forest (RF) achieved the highest accuracy with an $R^2$ of 0.9160 when optimized via Optuna.

Contributions

• Validates the use of machine learning for ETo estimation in semi-arid regions where meteorological data is limited.
• Establishes Optuna as a superior, more flexible, and computationally efficient alternative to traditional Grid Search for hydrological model optimization.

Funding

• Not specified in the provided text.

Citation

@article{Belarbi2025Efficient,
  author = {Belarbi, Zaid and Younoussi, Yacine El},
  title = {Efficient Hyperparameter Optimization for Reference Evapotranspiration Estimation with Limited Parameters: A Comparison of Optuna and Grid Search in the Doukkala Region, Morocco},
  journal = {Springer Link (Chiba Institute of Technology)},
  year = {2025},
  doi = {10.1051/e3sconf/202568000076/pdf},
  url = {https://doi.org/10.1051/e3sconf/202568000076/pdf}
}

Generated by BiblioAssistant using gemini-3-flash-preview (Google API)