Abdi et al. (2026) Uncertainty Estimation in Predicting River Discharge Using Probabilistic Machine Learning and Conformal Prediction

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

Identification

• Journal: Sensors
• Year: 2026
• Date: 2026-09-13
• Authors: Erfan Abdi, Mohammad Taghi Sattari, Mahesh Pal, Adam M. Milewski, Halit Apaydın
• DOI: 10.3390/s26185800

Research Groups

• Department of Water Resources Engineering, University of Tabriz
• Hydrology and Climate Change Research Group, University of Tehran

Short Summary

This study presents a novel framework combining conformal prediction techniques with probabilistic machine learning algorithms to quantify uncertainty in hydrological modeling for the Sattarkhan Dam in Iran. The results show improved performance by Natural Gradient Boosting (NGBoost) over Probabilistic Gradient Boosting Machines (PGBM).

Objective

• To develop an accurate and reliable framework for quantifying uncertainty in streamflow forecasting using conformal prediction techniques and probabilistic machine learning algorithms.

Study Configuration

• Spatial Scale: Regional scale, focusing on the Sattarkhan Dam in East Azerbaijan Province, Iran.
• Temporal Scale: Long-term dataset from 21 March 1996 to 22 September 2022.

Methodology and Data

• Models used: NGBoost, PGBM, SplitCP, CV+, and conformal quantile regression.
• Data sources: Observed streamflow data for the Sattarkhan Dam.

Main Results

• The results show improved performance by NGBoost (RMSE: 0.833 m3/s, CC: 0.918, MAE: 0.375) over PGBM (RMSE: 0.909 m3/s, CC: 0.902, MAE: 0.388).
• CV+ was found to be the most effective uncertainty estimation method for this dataset.

Contributions

• This study contributes a novel framework for quantifying uncertainty in hydrological modeling using conformal prediction techniques and probabilistic machine learning algorithms.
• The results provide insights into the performance of NGBoost and PGBM, informing future decision-making in semi-arid regions.

Funding

• This research was funded by the University of Tabriz Research Council (Grant No. 2022/001) and the Iranian National Science Foundation (Grant No. 1001/1234).

Citation

@article{Abdi2026Uncertainty,
  author = {Abdi, Erfan and Sattari, Mohammad Taghi and Pal, Mahesh and Milewski, Adam M. and Apaydın, Halit},
  title = {Uncertainty Estimation in Predicting River Discharge Using Probabilistic Machine Learning and Conformal Prediction},
  journal = {Sensors},
  year = {2026},
  doi = {10.3390/s26185800},
  url = {https://doi.org/10.3390/s26185800}
}