Abstract
This article explores the possibilities of forecasting water discharge in the Piskom River basin based on meteorological data using Machine Learning (ML) models. The study establishes relationships between the Piskom River flow and meteorological factors using Random Forest, XGBoost, and LSTM models, with their accuracy compared through various evaluation metrics (MAE, RMSE, R², and NSE). The analysis demonstrates that the Random Forest model provides the highest accuracy in forecasting the water discharge of the Piskom River. The research results indicate that ML models can serve as an effective tool for preliminary assessment of river flow and water resource management
Published
27-01-2026
Section
Articles
