Construction And Application Of Chinese Herbal Medicine Knowledge Map Based On Deep Learning

Chinese herbal medicine is the crystallization of the wisdom of countless ancestors for thousands of years and has been playing an irreplaceable role since ancient times.In Covid-19,Chinese herbal medicine with its unique advantages has played an important role in the prevention,treatment and subsequent recovery.Because most knowledge of Chinese herbal medicine exists in the form of books and literature,and there are problems of knowledge fragmentation and knowledge display singleness,most people do not have high recognition of Chinese herbal medicine.Therefore,in view of the massive data of Chinese herbal medicine,there is an urgent need for an effective data management method to integrate data and promote the promotion of Chinese herbal medicine.Knowledge graph can realize semantic link of network data and integrate and utilize multi-source heterogeneous data,so as to solve the problems of Chinese herbal medicine knowledge fragmentation and knowledge display.Therefore,this paper took Chinese herbal medicine as the research object,established Chinese herbal medicine data set,constructed Chinese herbal medicine knowledge map,designed and developed an intelligent service platform based on Chinese herbal medicine knowledge map,and accelerated the development of Chinese herbal medicine.The main contents of this paper are as follows:(1)Construct the data set of Chinese herbal medicine.By reading the literature of Chinese herbal medicine and discuss with the experts in the field,puts forward and defines the entities and entity relationship related to Chinese herbal medicine,and analyzes a variety of Chinese herbal medicine data characteristics of Chinese herbal medicine database records,to get the data,and with the tools of JIEBA participle word segmentation,duplicate data to eliminate and screening of data sample preprocessing operations.Because the collected data contains unstructured data,it cannot be used directly.Therefore,in order to obtain the data of Chinese herbal medicine in unstructured data.In this paper,Bi LSTM-CRF model was constructed to realize entity recognition of Chinese herbal medicine information.Experimental results showed that the accuracy rate of Chinese herbal medicine entity recognition model constructed in this paper reached 94.99%,recall rate reached 92.52%,F1 value reached 93.28,and the recognition effect was relatively ideal.After the extraction of the above methods,the data set of Chinese herbal medicine was finally established.A total of 8159 kinds of Chinese herbal medicines,25317 kinds of prescriptions,21625 kinds of usage and dosage information,8687 pieces of source information,9823 kinds of efficacy information and 20158 kinds of disease information were collected in this paper,providing data support for the construction of knowledge atlas of Chinese herbal medicines and the development of intelligent service platform in the next step.(2)Construct knowledge map of Chinese herbal medicine.a conceptual model of knowledge graph of Herbal medicine was constructed based on the top-down approach.First,with the help of medical workers and traditional Chinese medicine literature,the standard terms,concepts and relations of Chinese herbal medicine were summarized and analyzed,and the conceptual model of knowledge atlas of Chinese herbal medicine with Chinese herbal medicine as the core was designed.Secondly,the extracted data of Chinese herbal medicine was filled in according to the conceptual model.Finally,a statement was written by Py2 neo and imported into the Ne4 j graph database to complete the construction of the Knowledge graph of Chinese herbal medicine,which contained 6 types of entities with a scale of about 93,000.5 entity relationship categories,with a size of about 530,000.(3)Design and implement an intelligent service platform for knowledge atlas of Chinese herbal medicine.The main function of the system is knowledge question answering.Based on the knowledge atlas of Chinese herbal medicine,this paper chooses the method based on template matching to achieve question classification and parsing.In this paper,the Chinese herbal questions are divided into words by two-way maximum matching algorithm,and the key words are obtained,and the question template and query statement are created according to the key words matching to realize the question and answer function.It saves the time for users to obtain information and promotes the intelligent development of Chinese herbal medicine.This thesis around the knowledge of the Chinese herbal medicine,the knowledge system carried on the thorough research of Chinese herbal medicine,Chinese medicine data set and knowledge map construction and completed service platform for Chinese herbal medicine intelligence questions and answers function implementation,this is a practice of Chinese herbal medicine combined with Internet technology,to promote the development of Chinese herbal medicine,provide data support for the Chinese herbal medicine intelligence services.