Font Size: a A A

Design And Implementation Of The Server Side Of A Customer Service Robot System

Posted on:2017-01-11Degree:MasterType:Thesis
Country:ChinaCandidate:W J MaFull Text:PDF
GTID:2428330569985064Subject:Software engineering
Abstract/Summary:PDF Full Text Request
Customer service as the last barrier to the user experience,to undertake the work to protect the user experience: each business line and link of the enterprise to the user caused by the problems are required in the customer service link to solve and appease.With the increasing demand of new business enterprises,business complexity gradually increased,the number of users continues to rise,the customer service work is facing increasing pressure,and through a single telephone customer reception channels also can not caver all types of customer complaints and the scene.In order to reduce the pressure of artificial customer service,to make the diversity of the customer complaint to customer service representative and user,support user for help under different scenarios,we developed customer service robot.Online customer service is based on the accumulation of artificial customer service based on the common problems and answers to build customer service robot system,to support user input and provides fast answer.Customer service robot system in accordance with the software development process,First of all,based on the research background and research status,the system requirements are analyzed,hen the overall design of the system is carried out based on the requirements and the overall architecture diagram and execution flow chart,the whole system adopts C/S architecture,the server side uses C/C++ as the programming language,the database uses MySQL to store knowledge information,and through the E-R diagram and database table design.The server includes four modules,the control module is responsible for the demand distribution,the result sorting and the collection of user feedback;Customer service dialogue module is mainly through the comprehensive template matching,Bias classification,word vector technology and a variety of different models of spatial vector model to answer the match,the calculation of the weighted similarity score;Input problem guidance module is mainly based on user input,recommend the user problem,reduce user input costs,and guide users to clear intentions;Knowledge management module is mainly used to assist operators to manage and monitor the robot in order to better understand the user's state and optimize the robot.Finally,it tests the system from the function,performance and data level,and checks whether the system running in line with the expected standards.At present,the system recovery accuracy has reached 80%,the average customer service robot processing 2500 problems every day,the amount of processing has reached 1/3 of the artificial customer service.
Keywords/Search Tags:Customer service robot, Bayes classification, Word embedding, Vector Space Model
PDF Full Text Request
Related items