Design And Implementation Of Knowledge Graph Based Question Answering System In Maternal And Child Recipes

With the implementation of the universal two-child policy,the number of new-born babies is rapidly increasing,accompanied by the increasing demand for scientific diet for pregnant and newborn babies.However,there is a lack of systematic data in maternal and child recipes domain on the Internet,and the majority of the pregnant and their families cannot easily access accurate and scientific recipes.Therefore,using knowledge graph(KG)and natural language processing technology to develop a question answering(Q/A)system for the maternal and child recipes domain is not only in line with the new form of Internet+ promoted by the government but also can solve the social pain points.Based on the construction of KG and the implementation of Q/A algorithm,this paper finally implemented the Q/A system in maternal and child recipes.The main work of this paper includes:(1)The main work of constructing the knowledge graph in maternal and child recipes includes knowledge acquisition,knowledge extraction and knowledge storage.Among them,IDCNN-CRF model is used to automatically extract the acquired knowledge,and the Neo4j graph database is used to store the KG.(2)The main work of designing and implementing the Q/A algorithm includes semantic parsing and knowledge base retrieval.Among them,the BERT-BiLSTM-CRF model is used to complete the named entity recognition,and the BiLSTM combined with the cosine similarity calculation model is used to complete the attribute linking,so as to complete the semantic parsing of questions.After that,the natural language questions are transformed into database query language through the logic transformation algorithm to complete the knowledge base retrieval.(3)The main work of designing and implementing the KG based Q/A system in maternal and child recipes includes front-end display,backend logic processing and data storage,and the realization of both PC and mobile terminal.Among them,Flask,Vue.js and other web frameworks are used to integrate the above research content,so as to implement the Q/A system,and stress test is conducted on the system to verify its application ability.