In this article we provide an overview of data mining, informatics, and machine learning approaches for thermoelectrics. We describe how the initial development of a thermoelectric materials database has enabled the creation of a recommendation engine governed by machine learning and how this engine introduces a new paradigm in thermoelectric materials development. Performance probability is generated based on training models. A demonstration of the data mining approach is set forth in a ternary intermetallic system, where we report new materials.
Sparks, T. D., Gaultois, M. W., Oliynyk, A., Brgoch, J., & Meredig, B. (2016). Data mining our way to the next generation of thermoelectrics. Scripta Materialia, 111, 10-15. [invited paper]