The 79th JSAP Autumn Meeting, 2018

Presentation information

Symposium (Oral)

Symposium » Plasma Informatics - Development of Plasma Science by Taking Advantage of Big Data and Analytics

[19p-141-1~8] Plasma Informatics - Development of Plasma Science by Taking Advantage of Big Data and Analytics

Wed. Sep 19, 2018 1:30 PM - 5:15 PM 141 (141+142)

Takayuki Ohta(Meijo Univ.), Kazunori Koga(Kyushu Univ.)

3:00 PM - 3:30 PM

[19p-141-4] Descriptive, predictive and prescriptive problems on materials informatics

Hiori Kino1 (1.NIMS)

Keywords:materials informatics, functional ontology

We will explain three typical examples of materials informatics. The first one is the descriptive problem, where we evaluated the importance of descriptors of the Curie temperature. The second one is the predictive problem, where we increased the prediction accuracy by introducing the data segmentation to the potential energy surface given by the first-principles calculation. The third one is the prescriptive problem, where we employed Bayesian optimization in the structure search. We will also introduce an example of the functional decomposition tree to express and understand the knowledge the domain experts have.