Developing Infiltration Model: Random Forest for Micro-Hydro Power Planning DOI Open Access

Annisa R. Varhana,

Widya Utama, Rista Fitri Indriani

et al.

IOP Conference Series Earth and Environmental Science, Journal Year: 2024, Volume and Issue: 1418(1), P. 012055 - 012055

Published: Dec. 1, 2024

Abstract The goal of this study is to determine the classification infiltration for Micro-Hydro Power Planning using Random Forest (RF) machine learning algorithm. Utilizing Landsat 8 satellite imagery, data provides a comprehensive basis analyzing various environmental factors relevant infiltration. RF algorithm models and classifies rates, ensuring precise reliable predictions essential effective micro-hydro power planning. model evaluation results demonstrate excellent performance, with an Overall Accuracy 0.97 Kappa Coefficient 0.96, indicating strong agreement between predicted actual classifications. High Sensitivity, Specificity (0.99 all classes), User values (all above 0.95) underscore model’s ability correctly identify categories maintain consistency in positive negative predictions. Feature importance analysis highlights that certain spectral bands significantly enhance predictive capability, Band 3 playing crucial role (importance score 100), followed by Bands 7 6. These capture specific signatures associated different improving performance reliability. research contributes Sustainable Development Goals (SDGs), supporting SDG 6 (clean water sanitation), (affordable clean energy), 9 (industry, innovation, infrastructure), 13 (climate action), 15 (life on land) through improved resource management stewardship.

Language: Английский

Developing Infiltration Model: Random Forest for Micro-Hydro Power Planning DOI Open Access

Annisa R. Varhana,

Widya Utama, Rista Fitri Indriani

et al.

IOP Conference Series Earth and Environmental Science, Journal Year: 2024, Volume and Issue: 1418(1), P. 012055 - 012055

Published: Dec. 1, 2024

Abstract The goal of this study is to determine the classification infiltration for Micro-Hydro Power Planning using Random Forest (RF) machine learning algorithm. Utilizing Landsat 8 satellite imagery, data provides a comprehensive basis analyzing various environmental factors relevant infiltration. RF algorithm models and classifies rates, ensuring precise reliable predictions essential effective micro-hydro power planning. model evaluation results demonstrate excellent performance, with an Overall Accuracy 0.97 Kappa Coefficient 0.96, indicating strong agreement between predicted actual classifications. High Sensitivity, Specificity (0.99 all classes), User values (all above 0.95) underscore model’s ability correctly identify categories maintain consistency in positive negative predictions. Feature importance analysis highlights that certain spectral bands significantly enhance predictive capability, Band 3 playing crucial role (importance score 100), followed by Bands 7 6. These capture specific signatures associated different improving performance reliability. research contributes Sustainable Development Goals (SDGs), supporting SDG 6 (clean water sanitation), (affordable clean energy), 9 (industry, innovation, infrastructure), 13 (climate action), 15 (life on land) through improved resource management stewardship.

Language: Английский

Citations

0