Named Entity Recognition Method for CNC Machine Tool Design Knowledge Text
Hao Liu, Zheng Sun, Fangjian Ning · 2022 IEEE Conference on Telecommunications, Optics and Computer Science (TOCS) · 2022
This Aiming at the problems of unclear definition of terminology and lack of terminology recognition model in the current design knowledge text terminology recognition in the field of machine tool, a calibration strategy of machine tool design knowledge terminology is formulated, and a method of CNC machine tool design knowledge terminology based on bidirectional encoder representations from transformers is proposed. This method uses the idea of transfer learning, firstly performs fine-tuning based on the BERT pre-trained language model of the CNC machine tool field dataset, Then, the bidirectional long short-term memory (BiLSTM) network is used to extract the sequence feature information of the text, and finally terms are identified using conditional random field (CRF) calculations. The experimental results show that based on the BERT-BiLSTM-CRF model, the accurate recognition of the terms of CNC machine tool design knowledge is realized, which is more than 10% higher than the F1 of the existing recognition method.