基于在线评论的客户偏好趋势挖掘  

Customer preference trend mining based on online reviews

作　　者：沈超[1,2,3] 王安宁 陆效农[1,3] 彭张林 张强[1,3] Shen Chao;Wang Anning;Lu Xiaonong;Peng Zhanglin;Zhang Qiang(School of Management,Hefei University of Technology,Hefei 230009,China;School of Economics and Management,Anhui Polytechnic University,Wuhu 241000,China;Key Laboratory of Process Optimization and Intelligent Decision-making,Ministry of Education,Hefei 230009,China)

机构地区：[1]合肥工业大学管理学院,安徽合肥230009 [2]安徽工程大学经济与管理学院,安徽芜湖241000 [3]过程优化与智能决策教育部重点实验室,安徽合肥230009

出　　处：《系统工程学报》2021年第3期289-301,共13页Journal of Systems Engineering

基　　金：国家自然科学基金资助项目(71690235,72071060,71901086);安徽省科技重大专项项目(18030901028);安徽省自然科学基金项目(2008085QG336).

摘　　要：为了有效地从在线评论数据中获取客户的需求偏好,提出了一种客户偏好趋势挖掘方法.该方法采用时间序列模型预测下一阶段产品属性重要性,利用决策树模型分析客户偏好的变化趋势,将产品属性分为关键属性和非关键属性.并进一步,根据Mann-Kendall趋势检验将非关键属性分为过时属性、增值属性和稳定属性.此外,以汽车产品为案例,验证了该方法在产品设计与开发过程中起到的重要作用.研究结果可以为企业的汽车产品开发提供决策支持,从而使产品最大化地满足客户的需求.In order to effectively acquire customer demand preferences from online reviews data,this paper proposes a customer preference trend mining method.The time series model is used to predict the importance of product attributes of the next stage,and decision tree model is built to analyze the changing trend of customer preferences,by dividing product attributes into key attributes and non-critical attributes.Further,according to the Mann-Kendall trend test,non-critical attributes are classified into obsolete attributes,value-added attributes,and stable attributes.In addition,the automotive product is taken as an example to verify the important role of this method in product design and development.The research results can provide decision support for the development of automobile products,so that products can best meet customer needs.

关 键 词：在线评论 偏好趋势 产品属性 时间序列分析 决策树模型 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]