Flood Prediction Using a Stacked Ensemble of Random Forest, LSTM, and XGBoost (RFLE): A Multi-Basin Indian River Study

Priyanshu Raghav; Pallavi Joshi; Yatu Rani

Flood Prediction Using a Stacked Ensemble of Random Forest, LSTM, and XGBoost (RFLE): A Multi-Basin Indian River Study

Author(s)	Priyanshu Raghav, Dr. Pallavi Joshi, Dr. Yatu Rani
Country	India
Abstract	Flooding remains one of the most destructive and frequently occurring natural disasters worldwide, causing severe loss of life, property damage, and long-term displacement across riverine and coastal communities. This paper presents RFLE (Random Forest–LSTM–XGBoost Ensemble), a novel stacked ensemble flood prediction framework that combines three independently trained machine learning models through a Logistic Regression meta-learner using out-of-fold predictions. Random Forest captures complex non-linear feature interactions; the LSTM network models sequential temporal dynamics; and XGBoost applies gradient-boosted decision trees with explicit regularisation. A ten-year (2013–2022) daily multi-basin dataset spanning five major Indian river basins—Brahmaputra, Ganga-Yamuna, Godavari, Mahanadi, and Periyar—comprises 18,260 observations with 16 engineered hydrometeorological features derived from IMD gridded rainfall, ERA5 soil moisture, and GloFAS discharge data. The ensemble achieves an overall classification accuracy of 91.6%, binary AUC of 0.973, flood-class F1-score of 0.593, RMSE of 0.242, and MAE of 0.101 on the chronologically held-out 2022 test partition. An ablation study confirms that all three base models contribute non-redundant predictive information, with LSTM removal producing the largest single degradation.
Keywords	ensemble learning, flood prediction, GloFAS, LSTM, machine learning, multi-basin, Random Forest, time-series forecasting, XGBoost.
Field	Engineering
Published In	Volume 8, Issue 3, May-June 2026
Published On	2026-05-15

View / Download PDF File

E-ISSN 2582-2160

doi

CrossRef DOI is assigned to each research paper published in our journal.

IJFMR DOI prefix is
10.36948/ijfmr

Downloads

Research Paper Format Copyright Permission Form and Undertaking Form Cover Page Vol 8 Isu 3 Cover Page Vol 8 Isu 2 Cover Page Vol 8 Isu 2

All research papers published on this website are licensed under Creative Commons Attribution-ShareAlike 4.0 International License, and all rights belong to their respective authors/researchers.

CC-BY-SA

About IJFMR Fees & Payment Current Issue Publication Archive	Submit Research Paper Track Submission Status Publication Guidelines Publication Ethics Peer Review & Plagiarism	Join as a Reviewer Editors & Reviewers Reviewer Referral Program Get Reviewer Membership Certi.	Website/Journal Policies Usage Policy Content Policies Privacy Policy

Contact Us		+91-9687-828-838	editor@ijfmr.com

International Journal For Multidisciplinary Research

A Widely Indexed Open Access Peer Reviewed Multidisciplinary Bi-monthly Scholarly International Journal

Flood Prediction Using a Stacked Ensemble of Random Forest, LSTM, and XGBoost (RFLE): A Multi-Basin Indian River Study

Share this