An Ensemble Machine Learning Perspective on Rainfall Erosivity Dynamics Evaluation
Disaster Advances · 30.06.2026 · 33
DOI: 10.25303/198da33041 ↗Kurz gefasst
Fachartikel in der Zeitschrift Disaster Advances, online erschienen am 30.06.2026. Verfasst von Sachin Subhashrao Patil und 1 weiteren Autor:innen.
Abstract (Originalsprache)
Under tropical cyclones (TC) conditions, heavy storms that dump a lot of rain, often lead to serious soil erosion, especially in coastal regions which are already vulnerable. A small amount of storms leads to large share of annual erosive power. The traditional methods we have used like statistical models or physical models, do not always capture those really short bursts of intense rain very well. Therefore, to overcome such limitation, this study introduces machine learning based Ran-Boost, a hybrid ensemble model combining the designed improved Random Forest and extreme gradient boosting algorithms to improve erosivity estimation during tropical cyclone events. We designed synthetic dataset as algorithm which generates N samples by drawing values for wind speed, central pressure, sea surface temperature (SST) and land surface temperature (LST) and rainfall intensity randomly from their specified normal distributions. The proposed predictive framework known as RanBoost serves as an ensemble-based system specifically designed for tropical cyclone event rainfall erosivity analysis. We compared the performance of RF21 , XGBoost1 , Hybrid PSR-SVM-FFA20 and DNN14 with ensemble Ran-Boost model and result shows that proposed model has RMSE of 54.09 (lower than other) and R2 as 0.987 (higher than other).
Zitieren
APA
Patil, S. S., & Ridhorkar, S. (2026). An Ensemble Machine Learning Perspective on Rainfall Erosivity Dynamics Evaluation. Disaster Advances, 33. https://doi.org/10.25303/198da33041
Vancouver
Patil SS, Ridhorkar S. An Ensemble Machine Learning Perspective on Rainfall Erosivity Dynamics Evaluation. Disaster Advances. 2026:33. doi:10.25303/198da33041
BibTeX
@article{Patil2026,
author = {Patil, Sachin Subhashrao and Ridhorkar, Sonali},
title = {An Ensemble Machine Learning Perspective on Rainfall Erosivity Dynamics Evaluation},
journal = {Disaster Advances},
year = {2026},
pages = {33},
doi = {10.25303/198da33041}
}
EUVIMED