基于多流深度残差GRU网络的卷烟零售客户综合满意度评价模型  

Cigarette retail customer comprehensive satisfaction evaluation model based on multi-stream deep residual GRU network

作　　者：曾建新 刘奇燕 王德才 刘志 赵涛 ZENG Jian-xin;LIU Qi-yan;WANG De-cai;LIU Zhi;ZHAO Tao(Marketing center of China Tobacco Yunnan Industrial Co.,Ltd.,Kunming 650231,China)

机构地区：[1]云南中烟工业有限公司营销中心,云南昆明650231

出　　处：《云南民族大学学报（自然科学版）》2024年第4期513-520,共8页Journal of Yunnan Minzu University:Natural Sciences Edition

基　　金：云南中烟工业有限责任公司科技项目(2018QT03).

摘　　要：为全面精准掌握卷烟零售客户对烟草行业各区县分公司和相关单位在卷烟营销工作中的满意程度,不断完善客户服务方式,持续提升客户服务水平,提出一个基于多流深度残差GRU神经网络的卷烟零售客户综合满意度评价模型.该模型基于2301759条2015-2022年间的零售客户满意度调查问卷数据进行学习,将客户对服务的满意度划分为6个等级,并进行相应的情感分析,得到最终的情感分类结果作为综合满意度评价的依据.仿真对比实验和实际数据分析结果表明,所提模型具有较强的文本特征提取能力,相较传统的客户综合满意度评价方法可以获得更有效的情感分类效果和满意度评价结果,为相关行业零售市场的管理提供了新的视角和工具.In order to comprehensively and accurately grasp the satisfaction level of cigarette retail customers with each district and county branch in the tobacco industry in cigarette marketing work,continuously improve customer service methods and continuously enhance customer service,we proposed a cigarette retail customer comprehensive satisfaction evaluation model based on multi-stream deep residual GRU neural network.The model was based on 2300000 retail customer satisfaction questionnaire data during 2015-2022 for learning,classifying customer satisfaction with services into six levels and conducting corresponding sentiment analysis to get the final sentiment classification results as the basis for comprehensive satisfaction evaluation.Simulation comparison experiments and actual data analysis results showed that this model had strong text feature extraction ability and could obtain better sentiment classification results compared with traditional customer comprehensive satisfaction evaluation methods.It also provided new perspectives and tools for the management of retail markets in related industries.

关 键 词：深度学习 深度残差网络 GRU网络 卷烟零售 客户综合满意度评价 

分 类 号：TS4[农业科学—烟草工业]