基于SDAE与CART联合智能算法的通信网络用户满意度分析方法  

User Satisfaction Analysis Scheme of Communication Network Based on SDAE and CART Joint Intelligent Algorithm

作　　者：李露 于忠义[2] 李福昌[1] Li Lu;Yu Zhongyi;Li Fuchang(Wireless Technology Research Department,China Unicom Network Technology Research Institute,Beijing 100048,China;School of Mechanical Electronic&Information Engineering,China University of Mining and Technology,Beijing 100083,China)

机构地区：[1]中国联合网络通信有限公司网络技术研究院,北京100048 [2]中国矿业大学(北京)机电与信息工程学院,北京100083

出　　处：《信息通信技术》2020年第2期12-18,共7页Information and communications Technologies

摘　　要：论文提出一种基于栈式降噪自编码器(Stacked Denoising Autoencoder,SDAE)与分类和回归决策树(Classification and Regression Tree,CART)的移动互联网满意度预测方法,此模型能挖掘出用户的满意度与用户的特征和网络特征的关联规则,通过这种规则能更精准及时地预测到用户满意度的变化,以便运营商针对这种变化提前作出决策。论文所提方法能够挖掘特征间的深层关系,通过SDAE编码样本可以获得影响用户体验的隐含特征,及时发现用户对于网络贬损的真正痛点,为运营商网络建设和运行维护部门制定提升用户的网络感知策略提供依据,从而提升用户体验。A mobile internet satisfaction prediction method based on SDAE and CART decision tree is proposed.This method could mine the user's satisfaction with the user's characteristics and network characteristics association rules.Through rules,it could accurately and timely predict changes in user’s satisfaction,so that operators could make decisions in advance of such changes.The method proposed in the paper could mine the deep relationship between features.The SDAE code samples can be used to obtain the hidden features that affect user experience,discover the user's real pain points for network depreciation timely,and develop a network for the operator's network construction and operation.It could maintenance department to improve the user experience policies and provide the basis to improve the user experience.

关 键 词：栈式降噪自编码器 分类和回归决策树 人工智能 移动互联网 满意度 

分 类 号：TN929.5[电子电信—通信与信息系统] TP181[电子电信—信息与通信工程]