Research And Implementation Of Food Therapy Health Q&A Robot Base On Knowledge Graph

In recent years,people have begun to pay attention to how to adjust their bodies in daily life,and the topic of diet therapy is often mentioned because of this.The benefits of scientific dietary regimen for the body and human spirit are obvious.The most important thing is that it does not cause any burden on the body.Compared to medication,diet therapy is more about improving the physical fitness of the human body and playing a better preventive role.Secondly,historical materials such as food and fruits are more readily available in life and are more affordable and cheaper than drugs.Furthermore,diet therapy can improve one’s own body in a subtle way without making patients feel uncomfortable and painful.However,the current data of dietary health are still relatively discrete,and users cannot find a dietary solution for their own rehabilitation within an effective time.This article will focus on the construction of knowledge graph of diet therapy and the design of an intelligent question answering system:(1)To study how to construct a knowledge graph about dietary health,this paper will be divided into four aspects for detailed introduction.Knowledge extraction,mainly includes ontology-based extraction(including knowledge mining)and modelbased extraction.The data in this paper mainly come from structured and semistructured data of China Health Network and the medical information popularization Center,etc..The process is feature extraction,knowledge fusion,and supervised learning.Knowledge fusion mainly focuses on entity vector similarity calculation for entity disambiguation.Knowledge processing mainly focuses on knowledge reasoning(AMIE algorithm).This article implements knowledge storage by using Neo4 j and MySQL.(2)About the design of intelligent question answering,after the establishment of the diet knowledge graph,jieba is used to pre-process word segmentation and part-ofspeech tagging.Secondly,the Word2 vec model is used to calculate the similarity of the word vector,and then the feature vector is generated,and then learn feedback to users through naive Bayes pre-sorting and CNN sorting to more accurate answers.Finally,this article builds a knowledge spectrum of dietary health through the research of the above content,and improves the accuracy of the dietary question through various algorithms,and finally builds a dietary health question and answer system that meets user needs.