摘要
为了提高国网客服中心的自动问答能力,需要进行自动问答系统的相似度特征分辨计算,为此,提出一种基于平均互信息熵特征提取的国网客服中心自动问答系统相似度计算方法。根据深度学习和模糊聚类方法实现对面向客服的自动问答系统的相似度特征数据分类,通过数据分类结果进行针对性的客服问答。仿真结果表明,采用该方法进行面向客服的自动问答系统的相似度计算的准确性较高,数据分类性较好,提高了国网客服中心的客服自动问答效率。
In order to improve the automatic question-answering ability of the customer service center of the State Grid,it is necessary to carry out the similarity feature resolution calculation of the automatic question answering system.To this end,a method for calculating the similarity degree of the automatic question answering system of the national customer service center based on the average mutual information entropy feature extraction is proposed.According to the deep learning and fuzzy clustering methods,the similarity feature data classification of the customer-oriented automatic question answering system is realized,and the targeted customer service question and answer is obtained through the data classification result.The simulation results show that the similarity calculation of customer service oriented automatic question answering system is accurate and the data classification is good,which improves the efficiency of customer service automatic question answering in customer service center of China Network.
作者
吴佐平
刘迪
张千福
黄晓光
林鸿
WU Zuo-ping;LIU Di;ZHANG Qian-fu;HUANG Xiao-guang;LIN Hong(Beijing China-Power Information Technology Co.,Ltd.,Beijing 100085,China;State Grid Information&Telecommunication Group.,Ltd.,Beijing 102211,China)
出处
《信息技术》
2020年第3期99-103,共5页
Information Technology
关键词
自动问答
特征提取
大数据
automatic question and answer
feature extraction
big data
