A Hybrid Machine Learning Approach for Real-Time Beach Safety

Ensuring beach safety is vital for safeguarding public health and enabling recreational activities. This study introduces a hybrid framework that combines dynamic threshold-based risk filtering, unsupervised clustering, and ensemble classification to assess and categorize beaches based on safety levels. First, real‑time environmental and oceanographic measurements, such as wave height, water temperature, turbidity, and other meteorological conditions, are evaluated against adaptive thresholds: data points failing any criterion are immediately flagged as unsafe. Measurements that satisfy all thresholds are then processed through an unsupervised clustering algorithm to uncover distinct safety patterns. Finally, an ensemble of classification models is applied to clustered data to enhance predictive robustness and accuracy. When unsafe conditions are identified, the system alerts end users immediately.