Data-driven analysis on perovskite solar cell devices

Recent advancements in artificial intelligence (AI) techniques have significantly influenced daily life and the forefront of research and development. Data-driven research using AI accelerates the resolution of complex problems and aids in uncovering previously unknown knowledge and scientific discoveries. In this study, we propose a data-driven approach for investigating perovskite solar cells, a vibrant area within renewable energy applications. This approach incorporates the generation of a robust dataset, developing an interpretable machine learning model based on knowledge-based feature selection, and analyzing the impacts of material properties on the device performance. Through this framework, we successfully constructed accurate predictive models for the efficiency of perovskite solar cells and assessed the importance of each feature. Our analysis demonstrates that our models effectively capture existing knowledge about perovskite solar cells and can potentially inform the design of new perovskite solar cell configurations.