The Designation And Implementation Of Distributed Intelligent Question Answering System

users.At present,the intelligent question answering system for open Internet era is the era of information outbreak,Traditional call center based on artificial customer service,customer service center because of the existence of poor real-time,low normative,high cost and some other problems can not meet the needs of today's areas such as Siri and Cortana,although widely and flexibly applied in the field of open question and answer,can not meet the quiz needs in a limited area such as an industry or a company.So for the limited area of intelligent question and answer solutions has gradually become a research hotspot.This paper focuses on solving the background and requirements of the intelligent question answering in the limited area,and studies the relevant technical solutions needed in the realization of the intelligent question answering system.The deep learning solves the core functions of the question answering system and solves the problems of the traditional depth Learning model availability and performance problems in the production environment,the final design and development based on the completion of a distributed intelligent question answering system.The main research contents are as follows:1.To study how to use deep learning model to solve the problem of matching users with similar standards in knowledge base.In this paper,MC-BLSTM-MSCNN model based on convolutional neural network and recurrent neural network is used to solve the problem of user classification,and then the word vectors are used to solve the problem of matching similar problems;2.How to solve the problem of iterative training corpus annotation and low cost and high throughput storage.In this paper,the problem of annotation is solved by marking the session information unconsciously in the normal operation flow through the user and human customer service,and the storage solution is proposed by storing the iterative training corpus in Kafka;3.How to schedule dispatching delay task and Kafka cluster monitoring and alarming scheme distributed,and designs a distributed delay task scheduling framework based on Zookeeper and delay queue and a Kafka cluster monitoring and alarming system based on Kafka-manager and Zabbix.Based on the realization of the above key technical solutions,this dissertation completes the design and development of a distributed intelligent question answering system.The system aims at the question answering needs in the limited area,and realizes the user preprocessing,similar problem matching,quiz statistical analysis,knowledge base management and other functions,match the standard questions existing in the knowledge base according to the user questions and provide the corresponding answers,and provide the prerequisites for the call center and the call center to reduce the labor costs and improve the service quality.