Research On Text Classification Of TCM Nephropathy Based On Key Semantic Information

With the rapid development of artificial intelligence and the vigorous support of national policies for traditional Chinese medicine(TCM),the informationization and intellectualization of TCM have ushered in a new opportunity,which also have created a new opportunity to change the current dilemma of slow development of TCM.Based on the TCM disease text data provided by a Chinese medicine hospital,research of the TCM intelligent dialectics has been carried out.According to the characteristics of TCM disease text,a method of extracting the key semantic information of TCM disease text based on keywords is proposed.Then,based on a thorough understanding of the TCM syndrome differentiation,the dialectical problem of TCM nephropathy is abstracted into a supervised multi-classification problem,and deep learning is used to classify TCM nephropathy disease text.Finally,an intelligent service platform of TCM is constructed.The main contributions of this paper are as follows:1)Analysize the characteristics of TCM disease texts deeply,and domain words are recognized by using information entropy and other indicators of N-Gram fragments,which further improves the accuracy of word segmentation.2)Based on the characteristics of TCM disease text,a TF-IDF-DP algorithm based on disease location is proposed,which lays a foundation for obtaining the key semantic information of disease text.3)Based on the keywords of the disease text,this paper proposes a key semantic information extraction method based on keywords of TCM disease text(DKSIEK),which can extract the key semantic information of the disease position,symptoms,whether symptoms are related or not,and the severity of symptoms.The experimental results show that this method can not only effectively extract key semantic information,but also suppress noise.4)Investigate the text classification of deep learning deeply,and then applies it to the TCM nephropathy disease text classification comprehensively and systematically,and four deep learning classification models are given.Combining with the key semantic information,a classification method based on the fusion of the key semantic information of the disease text is proposed.At the same time,considering the characteristics of TCM syndrome differentiation,a classification method based on two stages is proposed.The experimental results show that the classification F1 value of the deep learning method is about 89%,and the improved two methods can effectively improve the classification F1 value by 2.6%and 3.6%,respectively.5)Investigate and analysize the existing service platform of TCM on the Internet deeply.Combining with the basis of laboratory about TCM,an intelligent service platform of TCM is designed and implemented,which is centered on intelligent dialectics of TCM and assisted by knowledge search,graph search and TCM question-and-answer.