Development of a web-based No coding machine learning platform for hydrology and environmental management - MoolML

초록

Developing data-driven models for hydrology and environmental management is challenging for non-experts, such as field engineers and environmental practitioners, due to limited coding experience and the complexity of model training and validation. To address this, we developed MoolML, a free, web-based, no-coding machine learning platform for simplified regression and classification modeling. The name MoolML is derived from the Korean word "(sic)" (mool), meaning "water," combined with Machine Learning (ML). MoolML integrates key functions such as data preprocessing, model training and prediction, hyperparameter tuning, cross-validation, feature importance analysis, and weather data collection, along with visualization tools for intuitive result presentation. The platform enables users to manage the entire modeling process without coding expertise while supporting data sharing and collaboration. The applicability and efficiency of developing ML models through the platform were tested using hydrological and environmental datasets from South Korea, and it is expected to support comprehensive watershed management.

키워드