Combining electronic and structural features in machine learning models to predict organic solar cells properties

Combining electronic and structural similarity between organic donors in kernel based machine learning methods allows to predict photovoltaic efficiencies reliably.