Extraction and Classification of TCM Medical Records Based on BERT and Bi-LSTM With Attention Mechanism

Traditional Chinese Medicine (TCM) medical records contain huge amounts of valuable medical information. However, in terms of text mining and utilization of TCM medical records, it is always difficult to extract and classify this information effectively. It is critical to identify a method of extracting and classifying the text from TCM medical records automatically. The method used in this paper attempts to apply a short medical record classification model based on BERT and Bi-LSTM with Attention mechanism. BERT prepossessing was used to obtain the short text vector as the input of the model. Result shows that the BERT-Bi-LSTM-Attention model achieves a highest average F1 value of 89.52% in the extraction and classification of TCM medical records, and therefore represents a significant improvement in modeling.