Research activities in the dielectric ceramics: information from data mining in literature

Vast published dielectric ceramics literature is a natural database for big-data analysis, discovering structure-property relationships and property prediction based on experimental conditions. We constructed a data-mining pipeline based on natural language processing (NLP) to extract property information from about 12900 dielectric ceramics articles. The micro-F1 scores for sentence classification, named entities recognition, relation extraction (related), and relation extraction (same) are 0.916, 0.827, 0.909, and 0.9, respectively. We built a dielectric ceramics database containing about 220000 aligned values of properties and normalized about 98000 data. Finally, we demonstrated the distribution of some important properties and the correlation relationships for different properties. We also analyzed the properties distribution for certain dielectric ceramics composites. The development of the dielectric ceramics was outlined. The experimental data enable us to extract the structure-property relationship in the future.