12:00 PM - 12:30 PM
[5a-S21-7] Experimental materials informatics by curation of thermoelectric properties on literature
Keywords:thermoelectric materials, materials informatics, database
Recently, materials informatics by first-principles calculations combined with machine learning is rapidly developing. However, inclusion of more experimental data is required for direct prediction of thermoelectric properties. In this study, we attempted to construct a database of experimental thermoelectric properties covering many plot data in published papers. We will introduce our processes for collection of 15 thousand papers, and the development of data-curation web system. We will also present our attempts to use the collected data, such as analysis and machine learning.