A Novel Technique for Rain/No-Rain Classification Using Outgoing Longwave Radiation Data from INSAT-3D

Designed a Random Forest rainfall-classification model on INSAT-3D satellite Outgoing Longwave Radiation (OLR) data, improving the Critical Success Index by 30.2% and the Equitable Threat Score by 30.9% over existing meteorological methods.

Addressed severe class imbalance with targeted SMOTE oversampling and random undersampling, improving recall by 56.1% while maintaining precision.