Design And Implementation Of Customer Service Chatbot For E-commerce

In recent years,online shopping has arisen,and artificial customer service posts have been born.Manual customer service plays an important role in pre-sales consultation,sales recommendation,after-sales return visit,dispute mediation and other shopping links.E-commerce companies spend huge labor costs on customer service.In order to reduce the cost of the company,the company raised the demand and established a set of ecommerce customer service robot system to assist the manual customer service.In recent years,chat robots mainly have two fields,one is open field and the other is for specific task areas.The e-commerce's manual customer work content is generally based on shopping-related issues.The customer service robot system for specific task areas is determined by demand analysis.The content of this thesis is to design and implement a customer service chatbot system for e-commerce.The system mainly consists of three modules: client,client side background management,and automatic question answering module.The automatic question answering module is the core of the system.This thesis first introduces the background and significance of researching customer service robots,and the research trends at home and abroad.Then,according to the business process analyzed by the requirements,the overall design of the system was carried out,and the client,the automatic question answering module and the customer service management platform were designed in detail.Next,implement the main module with code.The client implements the WeChat access function;the client management platform implements functions such as member management,role assignment,and customer tag configuration;the automatic question answering module implements model building,training,verification,etc.,using the word vector,the twin network,Key technologies such as convolutional neural networks.Finally,the test client,customer service management platform and Q&A module constitute a customer service robot system.In order to verify the function and performance of the designed system,a lot of experiments were carried out in this thesis.The experimental results show that: functionally,the customer service robot answers the questions,and the out-of-range transfer to the manual customer service can work normally;in terms of performance,the average response time of the detection system,the number of concurrent users and other indicators are in line with expectations,and the accuracy rate of the question and answer model meets the demand.