面向客户服务的聊天机器人系统设计

Design of Chat Robot System for Customer Service

本文设计了一种分层结构的聊天机器人系统,在互联网上搜集与顾客服务相关的对话数据,设计电商顾客服务领域的知识库,构建基于LSTM的Seq2Seq模型用于生成回答,使用朴素贝叶斯分类模型来对不同类型的问题选择合适的回答策略,并在必要时转向人工客服。

This paper designs a hierarchical chat robot system,which collects conversational data related to customer service on the Internet,designs a knowledge base in customer service field of e-commerce,builds Seq2Seq model based on LSTM to generate answers,uses Naive Bayesian classification model to select appropriate answer strategies for different types of questions,and turns to human service when necessary.