Bak et al. (2025) Development of a web-based No coding machine learning platform for hydrology and environmental management - MoolML

Identification

• Journal: Environmental Modelling & Software
• Year: 2025
• Date: 2025-12-10
• Authors: S. N. Bak, Jeongho Han, Gwanjae Lee, Nguyễn Đình Giang Nam, Joo Hyun Bae, Yeon Ho Jeong, Hyung-Jin Shin, Kyoung Jae Lim, Seoro Lee
• DOI: 10.1016/j.envsoft.2025.106830

Research Groups

• Department of Regional Infrastructure Engineering, Kangwon National University, Chuncheon-si, Gangwon-do, South Korea
• Agriculture and Life Sciences Research Institute, Kangwon National University, Chuncheon-si, Gangwon-do, South Korea
• ILEM Research Institute, Chuncheon-si, Gangwon-do, South Korea
• Rural Research Institute, Korea Rural Community Corporation, Ansan-si, Gyeonggi-do, South Korea
• Department of Biological and Agricultural Engineering, University of California, Davis, CA, USA

Short Summary

This paper introduces MoolML, a free, web-based, no-coding machine learning platform designed to simplify the development of regression and classification models for hydrology and environmental management, demonstrating its applicability and efficiency with South Korean datasets.

Objective

• To develop a user-friendly, web-based, no-coding machine learning platform (MoolML) that enables non-experts to build, train, and validate data-driven models for hydrological and environmental management.

Study Configuration

• Spatial Scale: Application tested using hydrological and environmental datasets from South Korea, focusing on watershed management.
• Temporal Scale: Not explicitly defined for the testing datasets, but the platform is designed for general application in long-term water sustainability and watershed management, implying various temporal scales depending on user data.

Methodology and Data

• Models used: MoolML platform, which integrates functions for developing various machine learning regression and classification models (e.g., hyperparameter tuning, cross-validation, feature importance analysis).
• Data sources: Hydrological and environmental datasets from South Korea, along with integrated weather data collection capabilities.

Main Results

• MoolML, a free, web-based, no-coding machine learning platform, was successfully developed, integrating data preprocessing, model training, prediction, hyperparameter tuning, cross-validation, feature importance analysis, and visualization tools.
• The platform enables users to manage the entire machine learning modeling process for regression and classification tasks without requiring coding expertise.
• The applicability and efficiency of developing ML models through MoolML were successfully tested using hydrological and environmental datasets from South Korea.

Contributions

• Development of the first free, web-based, no-coding machine learning platform specifically tailored for hydrology and environmental management, significantly lowering the barrier for non-expert practitioners.
• Integration of a comprehensive suite of ML functionalities (data preprocessing, model training, hyperparameter tuning, cross-validation, feature importance, visualization) into a single, intuitive platform.
• Facilitation of data sharing and collaboration among users, enhancing collective efforts in watershed management.

Funding

Not specified in the provided text.

Citation

@article{Bak2025Development,
  author = {Bak, S. N. and Han, Jeongho and Lee, Gwanjae and Nam, Nguyễn Đình Giang and Bae, Joo Hyun and Jeong, Yeon Ho and Shin, Hyung-Jin and Lim, Kyoung Jae and Lee, Seoro},
  title = {Development of a web-based No coding machine learning platform for hydrology and environmental management - MoolML},
  journal = {Environmental Modelling & Software},
  year = {2025},
  doi = {10.1016/j.envsoft.2025.106830},
  url = {https://doi.org/10.1016/j.envsoft.2025.106830}
}