Research On TCM Cases Segmentation And Symptom Matching Algorithm For TCM Diagnosis

On the road of intelligent development of TCM(traditional Chinese medical),the biggest obstacle is the problem of data.Therefore,it is necessary to transform more traditional empirical TCM data into quantifiable data and establish an effective and standardized database.Among them,the medical cases of TCM contain a large amount of basic diagnostic data,which should be fully utilized.To this end,in view of the problem of insufficient accuracy in the data processing methods of TCM cases,this paper proposes an attention-BIGRU-CRF model based on Bert word vectors for segmentation of TCM cases,and a weighted similarity algorithm for TCM Syndrome matching,so as to realize the extraction of syndromes in TCM cases.In addition,it implements TCM assisted diagnosis,and proposes a generalized priority syndrome differentiation algorithm and syndrome name fusion algorithm based on syndrome differentiation system.Use the syndrome extracted from the medical record to complete the auxiliary diagnosis of TCM.The contents and innovations are as follows:1.In order to solve the current problem of insufficient accuracy of word segmentation in TCM cases,this paper proposes an attention-BIGRU-CRF TCM case segmentation model based on the Bert word vector and provides a pre-trained segmentation model.The model structure is based on BIGRU-CRF,adding an attention mechanism and using Bert word vectors.On the open source dataset and the TCM case dataset,the segmentation results obtained are better than most of the segmentation models and Chinese segmentation tools.2.In order to achieve similar matching of medical case syndromes and diagnosis syndromes,this paper proposes a weighted algorithm of similarity of Chinese medicine syndromes.First,extract the syndromes based on the segmentation results,and prejudge the syndromes before calculating.Then use the cosine similarity algorithm based on TF-IDF and Bert word vector,and add the weight value to calculate the similarity.3.The syndrome differentiation system makes the diagnosis of TCM more standardized and scientific,and provides a good interface for the application of artificial intelligence and other computing sciences to TCM.Therefore,this paper implements the generalized priority syndrome differentiation algorithm and syndrome name fusion algorithm based on syndrome differentiation system,and with the help of the syndromes obtained from the TCM medical cases,the TCM assisted diagnosis is realized.