Presentation Information

[20a-P04-5]Examination of Automatic Magnetic Data Extraction from Scientific Documents

〇Masashi Ishii1, Hiroyuki Oka1 (1.NIMS)

Keywords:

TDM,NdFeB,Database

Efficient extraction of magnetic data from scientific documents was examined to perform materials informatics for magnet developments. The magnetic data were extracted primarily from tables using the machine extraction system previously created for polymer data. XML files of US patents were used as the scientific documents. The target magnetic properties were coercivity, remanence, maximum energy product, and squareness. In addition to these, the chemical compositions of samples were also extracted. At present, 2,457 data points for coercivity were extracted from 120 patents.