9:30 AM - 11:30 AM
[25a-P06-4] Sentence classification for polymer data extraction from scientific articles
Keywords:polymer data, sentence classification, materials informatics
Sentence classification in polymer articles was examined for mechanical extraction of polymer data. Eight classified labels were manually decided and the machine learning of sentence classification was performed. The training data was prepared with the labels and the 98 sentences in the sections of experimental, result, and discussion in polymer articles. The machine learning was performed using BERT and the accuracy was 0.74. The training data was small; however, the result was good. We will perform the learning with the enhanced data and report the results.