Research on the Automatic Punctuation Method of Sui and Tang Dynasty Epitaphs Incorporating Pinyin Information

Epitaphs are important historical, archaeological and research materials for ancient social life, however, there still exists a large amount of un-punctuated data, which undoubtedly hinders people from further understanding and exploring the deep information contained in these texts. Therefore, in this paper, we select the epitaphs of Sui and Tang dynasties as a specific research object, and construct a dataset using the existing punctuated texts, meanwhile, we innovatively propose an automated punctuation model GuwenBERT+ Pinyin+BiLSTM+CRF integrating the pinyin information, so as to achieve the deeper understanding of the text and accurate punctuation annotation. The experimental results show that the improved model improves the F1 value by 0.51% compared with the original model.