IMF2017

This presentation will provide an overview of computational strategies that harness statistical learning methods to selection and design of crystal chemistry for ferroelectrics and related classes of materials. We show how data driven methods can identify new parameters and correlations that augment theoretical and/or experimentally derived data. A particular emphasis of our discussion will be on the challenges to deal with the high dimensional nature of the data governing structure-chemistry-property relationships. Examples are provided in how one may use these methods to permit significant acceleration of materials discovery and design.