Developing machine learning models for accurate prediction of radiative efficiency of greenhouse gases

•The study presents a machine learning model using transfer learning to predict the radiative efficiencies of molecules.•A dataset of approximately 82,000 molecules was produced using a narrow band model and density functional theory (DFT) computed infrared spectra to facilitate radiative efficiency calculations.•The top-performing machine learning models were identified as DMPNN and ml-QM-DMPNN with feedforward neural networks.•The developed models correctly predict radiative efficiencies of molecules, contributing to the assessment of environmentally sustainable chemicals.