2018 年 61 巻 4 号 p. 200-205
In the field of material science the size of measured or calculated data is rather limited compared to the variety of materials, causing difficulty in material search with big-data type machine learning techniques. The present search system is designed to help material search by relating various material properties, enabling utilizing data from higher perspective viewpoint. Among properties in different categories such as thermal conductivity and electrical conductivity, relationship is extracted mainly from scientific principles. The relationships are made into a database and can be searched using algorithms in graph theory. In this paper, the framework, brief description of the relationship database, the graph expression of the relationships, and a relationship search using graph theory of the developed prototype system are presented. Examples are given to demonstrate the usefulness of the search system on material search.
