Dataset of ‘A Literature Mining Method of Fusing Text and Table Extraction in Materials Science’

We propose a named entity recognition model for material text, called SciBERT-Fasttext-BiLSTM-CRF (SFBC). We used this model to identify named entities from texts in the stainless steel scientific literature and shared data on the frequency of occurrence of selected entities in this database between 2012 and 2021. By analysing the data in this dataset, researchers are able to understand the top research trends in stainless steel materials over the last decade.

我们提出了一种面向材料文本的命名实体识别（Named Entity Recognition, NER）模型，命名为SciBERT-Fasttext-BiLSTM-CRF（SFBC）。我们借助该模型从不锈钢领域的科技文献文本中识别命名实体，并共享了2012至2021年间该数据库内选定实体的出现频次数据。研究人员可通过分析该数据集内的数据，明晰近十年来不锈钢材料领域的主流研究趋势。