Traditional Chinese medicine prescription recommendation method based on symptom term mapping and deep learning

Abstract

The invention provides a traditional Chinese medicine prescription recommendation method based on symptom term mapping and deep learning, and belongs to the technical field of traditional Chinese medicine prescription recommendation. All symptom words of a patient are subjected to symptom term mapping, network embedding representation, symptom feature fusion, convolutional neural network learning, full-connection network training and activation layer activation; obtaining a prediction probability of each candidate traditional Chinese medicine; the prediction probability of each traditional Chinese medicine is arranged in a descending order, and the corresponding traditional Chinese medicine sequence after descending is used as the final recommendation sequence. According to the invention, traditional Chinese medicine prescription recommendation is carried out based on symptom term mapping and deep learning, the'out-of-word 'problem of clinical symptoms is relieved through symptom term mapping, prescription recommendation is carried out in combination with deep learning, and an attention-based prescription recommendation feature contribution degree evaluation module is established. A feature fusion-based symptom term representation and patient clinical feature fusion strategy is established, and the reliability of prescription recommendation is improved.

Description

technical field [0001] The invention relates to the field of medical technology, in particular to a TCM prescription recommendation method based on symptom term mapping and deep learning. Background technique [0002] In recent years, a large amount of TCM prescription data formed in the course of clinical practice has not been fully utilized. Combining existing TCM clinical prescription data with artificial intelligence-related methods to recommend TCM intelligent prescriptions can play an auxiliary role in doctors' diagnosis and treatment. [0003] Electronic medical records of clinical patients usually contain information such as patient complaints, current medical history, and treatment prescriptions. The recording of patient phenotype information in such data is usually highly subjective. TCM intelligent prescription recommendation refers to the use of artificial intelligence, data mining and other computer technologies to build a prescription recommendation model throu...

AI Technical Summary

Problems solved by technology

[0006] Although the existing work on prescription recommendation has achieved certain results in mining and utilizing electronic medical records of traditional Chinese medicine, on the whole, the existing prescription recommendation methods have failed to achieve a good performance in recommendation results; in addition, many clinical manifestations of traditional Chinese medicine There is a synonymous relationship among the phenotypes, and most of the terms that have not appeared before can be obtained based on the combination of words in the existing phenotype terms. For example, the term "foot pain" has a synonymous relationship with the term "foot pain", But both are considered as two different characteristics

At the same time, for the "unregistered words" that may appear in the prescription recommendation, that is, the symptom words that do not appear in the terminology database, the existing recommendation methods cannot effectively solve the problem of how to use existing knowledge to represent the "unregistered words"

Images