Research On Semantic Analysis In Pharmaceutical Field Based On Knowledge Graph

With the advent of the Internet era,healthcare services have entered a new era of Internet Plus Healthcare,which integrates information technology such as mobile communication,cloud computing,Internet of Things,and big data with traditional healthcare services.However,the influx of complex data brought by the Internet has also increased the difficulty of information retrieval and learning for people.In the field of healthcare,medical knowledge graph serves as the foundation for implementing medical artificial intelligence,and the construction of a comprehensive medical knowledge graph can provide people with more efficient and accurate healthcare services.This thesis focuses on the research of semantic analysis in the pharmaceutical domain,which includes the following aspects:(1)Research on how to build a high-quality and effective knowledge base for pharmaceutical domain question-answering systems based on knowledge graph technology.The main steps include knowledge acquisition,data preprocessing,entity recognition,relation extraction,and knowledge storage.(2)Research on semantic analysis based on semantic networks.This thesis applies the GBDT + LR multi-model fusion algorithm and the Bert + Text CNN multi-intent classification algorithm for secondary intent recognition.In addition,a named entity recognition model based on Bi LSTM-CRF is used to accurately extract entities from the text,and AC automaton is further employed for refinement.This model combines the advantages of Bi LSTM and CRF,enabling effective identification of named entities in pharmaceutical texts.(3)Building a front-end interface using Uniapp and a back-end service using Sanic to provide accurate answers to healthcare seekers through online Q&A.The research aim is to achieve effective semantic analysis of text information in the pharmaceutical domain by combining knowledge graph technology and natural language processing techniques.The research results demonstrate that this approach can effectively extract valuable information from pharmaceutical texts and provide a powerful tool for knowledge management and decision support in the pharmaceutical domain.Through the aforementioned methods,this thesis successfully constructs a Chinese medical knowledge graph presented in the Neo4 j graph database.Moreover,by utilizing deep learning methods,the intelligent question-answering system based on the medical knowledge graph has been enhanced in understanding user’s natural language questions and intents.Ultimately,an online healthcare platform that meets user needs has been achieved.