A remote supervision method for fine-grained syndrome name segmentation in traditional Chinese medicine

Abstract

The invention relates to a remote monitoring method for realizing fine-grained syndrome name segmentation of traditional Chinese medicine. The method uses the naming and translation conventions of TCMsyndromes to train a conditional random field model based on the segmentation of TCM fine-grained syndromes names, Firstly, the Chinese-English word relationship of TCM syndromes in Chinese-English parallel corpus is used to construct the noisy marker training data by probability alignment method, which can heuristically combine forward alignment and backward examination to infer the fine-grainedsegments in TCM syndromes, and then the conditional random field model of these fine-grained segments is trained. The invention is reasonable in conception and simple in use for training a conditional random field model based on fine-grained syndrome name segmentation of traditional Chinese medicine in the field of traditional Chinese medicine, and hardly relies on or even does not need manual labeling training data.

Description

technical field [0001] The invention belongs to the medical field, and in particular relates to a remote supervision method for fine-grained syndrome name segmentation in traditional Chinese medicine. Background technique [0002] Syndrome is a concept unique to Traditional Chinese Medicine (TCM). It is also an important pillar of Traditional Chinese Medicine (TCM). The name of the syndrome includes the etiology, the location of the disease, the nature of the disease, the tendency of the disease, the pathology, and the relationship between the healthy Qi and the disease-causing Qi. A clear understanding of TCM syndromes helps researchers discover the laws of TCM. It is generally recognized that a clear understanding of the syndrome will help researchers discover the laws of TCM and help integrate TCM and Western medicine. Syndromes are usually presented in a coarse-grained form, however, the fine-grained medical information underlying coarse-grained TCM syndromes has not ...

AI Technical Summary

Problems solved by technology

[0006] In order to solve the problem in the above-mentioned background technology that the segmentation of coarse-grained symptoms into fine-grained forms requires professionals to manually mark data, the present invention proposes a remote supervision method for fine-grained syndrome name segmentation in traditional Chinese medicine, which hardly relies on or even requires Manually mark the training data, the feasibility and effectiveness have been verified by experiments, and the highest F1 score reaches 0.9177, which is a competitive value compared with the CRF result (0.939) trained by manually splitting the data

Images