Marine Shellfish Entity Recognition Based on BiLSTM-CRF Model

Named entity recognition is an important research area in natural language processing. For marine shellfish biological identification tasks, this article builds a Bilstm-CRF model for physical identification tasks. The data set used in the training is a self-built data set MSDataset.The experimental results are:After 50 rounds of training, the average F1 value of the various entity recognition results was 89.75%, which was 18% and 5.5% higher than the 71.75% of HMM and 84.25% of CRF respectively, indicating that the entity recognition could be effectively improved after adding the BiLSTM network layer.Meanwhile, the overall results of accuracy and recall of the BiLSTM-CRF model were better than those of the HMM and CRF models. Therefore, the BiLSTM-CRF model can effectively achieve the entity recognition of marine shellfish organisms, and effectively improve the entity recognition accuracy of marine shellfish organisms.