Samantaray et al. (2026) Runoff Prediction Based on Support Vector Machine Optimized with Arithmetic Optimization Algorithm

Identification

• Journal: Lecture notes in civil engineering
• Year: 2026
• Date: 2026-01-01
• Authors: Sandeep Samantaray, Abinash Sahoo, Deba P. Satapathy
• DOI: 10.1007/978-981-95-0736-8_14

Research Groups

• Department of Civil Engineering, NIT Srinagar, Srinagar, Jammu & Kasmir, India
• Department of Civil Engineering, OUTR Bhubaneswar, Bhubaneswar, Odisha, India

Short Summary

This paper develops a novel hybrid model, AOA-ELM, by optimizing the Extreme Learning Machine (ELM) with the Arithmetic Optimization Algorithm (AOA) for runoff prediction. The study demonstrates that AOA-ELM significantly outperforms the standalone ELM model in terms of R², NSE, and RMSE for hydrologic time-series prediction in the Subarnarekha River basin.

Objective

• To develop and evaluate a hybrid model (AOA-ELM) for accurate runoff prediction to support water resource planning, disaster reduction, and flood control, addressing the challenges of uncertainties in runoff forecasting.

Study Configuration

• Spatial Scale: Subarnarekha River basin, India
• Temporal Scale: Data from 2004 to 2023 (20 years), with 2004–2018 used for training and 2019–2023 for prediction.

Methodology and Data

• Models used: Hybrid AOA-ELM (Arithmetic Optimization Algorithm - Extreme Learning Machine) and conventional ELM (Extreme Learning Machine) for comparison.
• Data sources: Historical runoff data for the Subarnarekha River.

Main Results

• The proposed AOA-ELM model consistently outperformed the standalone ELM model across all evaluated performance metrics, including the coefficient of determination (R²), Nash–Sutcliffe efficiency (NSE), and root mean square error (RMSE).
• The Arithmetic Optimization Algorithm (AOA) proved to be an effective training approach for selecting and optimizing ELM parameters.
• AOA-ELM significantly enhanced the generalization performance of ELM for hydrologic time-series prediction.
• The developed model is presented as a vital resource for accurate flood forecasting and prompt alerts.

Contributions

• Introduction of a novel hybrid AOA-ELM model, integrating the Arithmetic Optimization Algorithm with Extreme Learning Machine for improved runoff prediction.
• Demonstration of the effectiveness of AOA as a training approach to enhance the generalization performance of ELM in hydrological time-series forecasting.
• Provision of a robust and accurate model that can serve as a valuable tool for flood forecasting and early warning systems.

Funding

• Not explicitly mentioned in the provided text.

Citation

@article{Samantaray2026Runoff,
  author = {Samantaray, Sandeep and Sahoo, Abinash and Satapathy, Deba P.},
  title = {Runoff Prediction Based on Support Vector Machine Optimized with Arithmetic Optimization Algorithm},
  journal = {Lecture notes in civil engineering},
  year = {2026},
  doi = {10.1007/978-981-95-0736-8_14},
  url = {https://doi.org/10.1007/978-981-95-0736-8_14}
}