<![CDATA[Atmospheric physics is very useful in predicting rainfall, particularly for analyzing air saturation conditions as a prerequisite for condensation. This study aims to model rainfall prediction using thermodynamic parameters, namely relative humidity (RH) and dew point temperature difference (ΔT). These parameters were collected from BMKG Lampung meteorological data (2022–2024) and processed using the Magnus equation. ΔT is important as a sensitive indicator of air unsaturation. The data were statistically analyzed and modeled using a Gradient Boosting Classifier. The results obtained indicate a strong correlation between RH and ΔT and rainfall events (point-biserial correlation of 0.475). Furthermore, ΔT during rainfall is lower (average 2.87°C) and stable, indicating near-saturation conditions. During the evaluation stage, the model achieved 76% accuracy and 84% recall during rainfall. The model's good performance proves the effectiveness of physical parameters as predictive features. Finally, the model was implemented in a Flask-based web application for practical accessibility.]]>
Copyrights © 2025
