网络环境下考虑目标顾客偏好的服务产品选择方法  

A Method for Selecting Service Products Considering Target Customer’s Preference in the Network Environment

作　　者：郭晓春[1] 马玉梅 曹萍萍[1,2] GUO Xiaochun;MA Yumei;CAO Pingping(Department of Basic Teaching and Research,Criminal Investigation Police University of China,Shenyang 110854;Business School,Liaoning University,Shenyang 110136)

机构地区：[1]中国刑事警察学院基础教研部,沈阳110854 [2]辽宁大学商学院,沈阳110136

出　　处：《系统科学与数学》2022年第7期1769-1787,共19页Journal of Systems Science and Mathematical Sciences

基　　金：公安部技术研究计划基科费项目(2021JSYJC23);辽宁省社会科学规划基金项目(L21BGL056);辽宁省教育厅科学研究经费项目(LJKZ0075);中央高校基本科研业务费重大项目培育计划(D2021009)资助课题。

摘　　要：随着互联网技术的发展,各类电商平台应运而生,所产生的在线评论有助于顾客在购买前更好地了解服务产品.然而,在线评论数量的巨大也为顾客带来负担.为了帮助目标顾客更好地借助在线评论从众多的同类型服务产品中找出符合其偏好的服务产品,文章提出一种考虑目标顾客偏好的服务产品选择方法.首先,根据相关电商平台上的在线评分和文本评论,采用K-Means算法确定服务产品的评价维度;其次,基于概率分布理论,确定基于在线评分的备选服务产品针对各评价标度的概率分布;然后,采用卷积神经网络对文本评论进行情感分类,进而确定基于文本评论的备选服务产品针对各评价标度的概率分布;进一步地,构建总体规范化评价值矩阵;之后,考虑目标顾客关于评价维度的偏好,确定评价维度的相对权重,并运用TODIM方法获得备选服务产品的排序结果;最后,以携程网上的在线评论为依据进行实例研究,说明了该方法的可用性.With the development of Internet technology,various e-commerce platforms have emerged,and online reviews generated can help customers better understand service products before purchasing.However,the sheer volume of online reviews also places a burden on customers.In order to help target customers better find out the service products that meet their preferences from numerous service products of the same type with the help of online reviews,this paper proposes a service product selection method considering the target customer’s preference.Firstly,according to online ratings and text reviews on relevant e-commerce platforms,the evaluation dimension of service products is determined by using the K-Means algorithm.Secondly,based on the probability distribution theory,the probability distribution of each alternative service product based on online ratings for each evaluation scale is determined.Thirdly,the convolutional neural network is used to classify the sentiment of text reviews,and then the probability distribution of each alternative service product based on text reviews for each evaluation scale is determined.Further,the overall normalized evaluation matrix is constructed.Then,the relative weights of evaluation dimensions are determined considering the preferences of target customers on evaluation dimensions,and the ranking results of the alternative service products are obtained by using the TODIM method.Finally,a case study based on online reviews on Ctrip is given to illustrate the usability of the method.

关 键 词：服务产品选择 目标顾客偏好 网络环境 K-MEANS算法 卷积神经网络 

分 类 号：F274[经济管理—企业管理] F724.6[经济管理—国民经济] TP183[自动化与计算机技术—控制理论与控制工程] TP391.1[自动化与计算机技术—控制科学与工程]