Prediction of seebeck coefficient for compounds without restriction to fixed stoichiometry: A machine learning approach

摘要: The regression model-based tool is developed for predicting the Seebeck coefficient of crystalline materials in temperature range from 300 K to 1000 K. accounts single crystal versus polycrystalline nature compound, production method, and properties constituent elements chemical formula. We introduce new descriptive features relevant prediction coefficient. To address off-stoichiometry materials, predictive trained on a mix stoichiometric nonstoichiometric materials. implemented into web application (http://info.eecs.northwestern.edu/SeebeckCoefficientPredictor) assist field scientists discovery novel thermoelectric © 2017 Wiley Periodicals, Inc.