Research on Named Entity Recognition Based on ELECTRA and Intelligent Face Image Processing
Yihui Fu, Fanliang Bu · 2021 IEEE International Conference on Emergency Science and Information Technology (ICESIT) · 2021
Aiming at the problem that the corpus of drug-related fields is not rich and the relevant information of drug-related personnel is insufficient, this paper constructs a 600,000-word-scale drug-related text data set, and proposes a named entity recognition method for drug-related personnel based on ELECTRA-BiLSTM-CRF. First input the labeled text into the ELECTRA pre-training language model to obtain a word vector with better semantic representation; then input the trained word vector into the bidirectional long short-term memory (BiLSTM) network to extract the context feature; finally, the best predicted label sequence is obtained through the conditional random field(CRF). The performance of this model was evaluated on the drug-related text data set. The experimental results showed that the F1 value of the ELECTRA-BiLSTM-CRF model reached 94%, which was better than the BERT-BiLSTM-CRF, BERT-CRF, and BiLSTM-CRF models, which proved this model has a good effect on the named entity recognition of drug-related personnel.