Multi-Task Joint Learning Model for Chinese Word Segmentation and Syndrome Differentiation in Traditional Chinese Medicine

Abstract

Evidence-based treatment is the basis of traditional Chinese medicine (TCM), and the accurate differentiation of syndromes is important for treatment in this context. The automatic differentiation of syndromes of unstructured medical records requires two important steps: Chinese word segmentation and text classification. Due to the ambiguity of the Chinese language and the peculiarities of syndrome differentiation, these tasks pose a daunting challenge. We use text classification to model syndrome differentiation for TCM, and use multi-task learning (MTL) and deep learning to accomplish the two challenging tasks of Chinese word segmentation and syndrome differentiation. Two classic deep neural networks—bidirectional long short-term memory (Bi-LSTM) and text-based convolutional neural networks (TextCNN)—are fused into MTL to simultaneously carry out these two tasks. We used our proposed method to conduct a large number of comparative experiments. The experimental comparisons showed that it was superior to other methods on both tasks. Our model yielded values of accuracy, specificity, and sensitivity of 0.93, 0.94, and 0.90, and 0.80, 0.82, and 0.78 on the Chinese word segmentation task and the syndrome differentiation task, respectively. Moreover, statistical analyses showed that the accuracies of the non-joint and joint models were both within the 95% confidence interval, with pvalue < 0.05. The experimental comparison showed that our method is superior to prevalent methods on both tasks. The work here can help modernize TCM through intelligent differentiation.

Efficacy of Xinbao pill on chronic heart failure: Study protocol of a multicenter, randomized, double-blind, placebo-controlled trial

“…First, it is difficult to conduct clinical studies under the guidance of the "syndrome differentiation and treatment" theory of TCM (Jiang et al, 2012). The principle of TCM compounding is to treat patients based on different clinical manifestations of diseases under the guidance of TCM theory, which is summarized as "syndrome," and then to administer the corresponding Chinese medicine treatment (Hu et al, 2022). According to the theory of TCM, XBP mainly treats CHF patients with "heart-kidney yang deficiency syndrome."…”

“…First, it is difficult to conduct clinical studies under the guidance of the “syndrome differentiation and treatment” theory of TCM( Jiang et al, 2012 ). The principle of TCM compounding is to treat patients based on different clinical manifestations of diseases under the guidance of TCM theory, which is summarized as “syndrome,” and then to administer the corresponding Chinese medicine treatment ( Hu et al, 2022 ). According to the theory of TCM, XBP mainly treats CHF patients with “heart-kidney yang deficiency syndrome.” However, in actual clinical practice, patients present with different symptoms so the treatment approach is subjective based on the doctor’s belief, and the same patient may even receive a different diagnosis from a different doctor ( Hao et al, 1990 ).…”

Research on TCM syndrome differentiation based on multi-feature fusion and GCN

“…SDTM uses LDA (Latent Dirichlet Allocation) to capture the relationship between symptoms and syndrome types, and combines similarity to measure the relationship between topics and syndrome types. Hu et al [32] proposed a method of TCM Yin deficiency and Yang deficiency syndrome differentiation based on BiLSTM and CNN. In literature [32], BiLSTM was used for Chinese word segmentation task and CNN was used for TCM syndrome differentiation task.…”

Sentiment Classification of Chinese Tourism Reviews Based on ERNIE-Gram+GCN

“…Most of the current text classification methods combine models such as CNN and LSTM [ 15 , 16 , 17 ], which prioritize sequential and local information and can capture semantic and syntactic information in continuous word sequences very well. However, they ignore global word co-occurrence, which carries discontinuous as well as long-distance semantic information.…”