¹ Department of Physics, University School of Sciences, Rayat Bahra University, Mohali, Punjab 140104, India.

* Correspondence: reetasharma2012@gmail.com

Machine learning provides a powerful approach for predicting material properties from existing experimental data. In this study, the optical band gap of In–Se chalcogenide thin films was predicted using a literature-derived dataset containing film thickness, indium composition, and selenium composition. A Random Forest regression model was developed to capture the nonlinear relationship between these parameters and the optical band gap. The model demonstrates strong predictive performance with a coefficient of determination (R² = 0.929) and a root mean square error (RMSE = 0.071 eV), indicating close agreement between predicted and reported band gap values. Feature-importance analysis reveals that film thickness is the dominant predictor, contributing about 63% of the total model importance, followed by selenium and indium composition. The observed dependence reflects underlying physical effects such as thickness-related electronic confinement and compositional variations that influence the electronic structure of In–Se thin films. These results demonstrate that literature-driven machine learning can provide rapid and reliable estimation of optical band gaps, offering an efficient strategy for guiding the design of chalcogenide thin films with targeted optoelectronic properties.