Abstract
This paper independently explores the chapter structure of scientific and technological literature in the field of shipbuilding in the natural sciences and the field of library and information in the social sciences. The chapter structure model of previous studies, namely 'background, purpose, method, result, conclusion, demonstration,' is quoted as the verification object of the document chapter structure in the field of exploration. In order to verify the rationality of the structure, this paper uses the deep learning models TextCNN, DPCNN, TextRCNN, and BiLSTM-Attention as experimental tools, and designs 5-fold cross-validation experiment and normal experiment, and finally verifies the rationality of the model structure, and It is concluded that the BiLSTM-Attention model can better identify the chapter structure in this field.
Recommended Citation
Ma, Huan; Fu, Zhu; and Jia, Qiong, "Automatic Recognition of Knowledge Characteristics of Scientific and Technological Literature from the Perspective of Text Structure" (2021). ICEB 2021 Proceedings (Nanjing, China). 63.
https://aisel.aisnet.org/iceb2021/63