Research And Implementation Of Intelligent QA Enhancement System For Vertical Domain

Due to the development of artificial intelligence technology and the maturity of natural language processing technology,natural language processing technology has derived many applications,the most common of which is intelligent question answering system.Intelligent question answering(QA)system has been successfully applied to enterprises to help them form intelligent solutions.The intelligent question answering system allows users to talk with the system in understandable natural language.The system processes the language information input by users and gives the answers that meet the needs of users.In the intelligent question answering system,the core is the understanding and processing of user language,which is a long-term optimization process.Based on the existing intelli gent question answering system of the company,this paper optimizes the intelligent question answering dialogue module based on knowledge base,uses natural language processing technology to improve the effect of question answering dialogue,and improves the recall rate of knowledge points and the accuracy rate of the model.This paper is divided into the following three parts:1.Design the effect optimization function point of intelligent QA dialogue based on knowledge base.The basis of question answering module is to generate knowledge points from the original corpus information and add them to the knowledge base,and configure corresponding answer information for each knowledge point to complete the automatic response.On the basis of the original question and answer,in view of the existing problems,the factors that affect the effect of intelligent question and answer dialogue are designed with corresponding functions.The factors that affect the effect of question and answer lie in the complexity and diversity of the user's language information.How to generalize the semantic expression of the same semantic and different forms and improve the accuracy of question and answer are the problems to be solved in this paper.Common problems are similar questions of knowledge points and confusion ofknowledge points.By calculating the text similarity between knowledge points and the distance between core words to optimize.At the same time,data analysis is used to visualize the data indicators on the platform.Use Python to monitor the effect of visual intelligent QA dialogue.Through different forms of charts to show the use of intelligent QA and core indicators2.Study the word vector technology.Through the introduction of word vector processing technology,combined with Jieba word segmentation to preprocess the corpus data,as well as to quantify the text,use the cbow algorithm model in word2vec to train the feature words,and use the algorithm to calculate the text similarity to learn the similar questions in the intelligent QA dialogue3.Improve the effect of intelligent QA dialogue by optimizing the process of building knowledge base.For the generation of knowledge points in the knowledge base,cluster mining method is used to mine valuable corpus information from the corpus to form several corpus clusters,from which valuable knowledge points are formed.Data mining can mine more valuable data information to support the use of the project and improve the effect of QA4.Using Python development language and MySQL database,using Django web framework,using Vue framework to develop the front end,using node.js as the middle layer,based on the original QA system,an intelligent QA enhancement system is designed to enhance the effect of QAAfter three months of iterative development,the intelligent QA enhancement system has been successfully launched through functional test and performance test,And in 5 online projects completed landing.the recall rate of intelligent QA significantly increased by 4.2%,The accuracy of QA increased by 2.1%.