Unsupervised natural language processing and machine learning tools for energy material screening

Discovery of new materials in energy application has drawn attention of the scientist in last few years. Hence, still there are promising existent materials with unknown capabilities. Extracting regarded materials, requires enormous amount of effort and time, to go through millions of literature. Advancements in Natural Language Processing algorithms made it feasible to obtain embedded data in literature in unsupervised manners. Superconductivity has been the focus of enormous research effort since its discovery more than a century ago. Yet, some features of this unique phenomenon remain poorly understood; prime among these is the connection between superconductivity and chemical/structural properties of materials. To bridge the gap, several machine learning schemes are developed herein to model the critical temperatures (Tc) of roughly 30000 known superconductors available via the SuperCon database. The model uses coarse-grained features based only on the chemical compositions. It shows strong predictive power, with out-of-sample accuracy of about 92%. Separate regression models are developed to predict the Tc values of high-T and low-T compounds achieved from NLP technique applied on literature abstracts and fed to scraping algorithm for novelty check.