现今，网络购物依托互联网的发展已成为社会普遍现象。网络零售顾客逐年 增加，网络零售额占全国社会消费品零售总额的比重不断攀升。网络零售平台上 充斥着大量的产品评价数据，这些数据直接反应了顾客满意。近年来对数据资源 的利用效率影响着企业的发展。顾客满意决定着未来网络零售市场的发展和网络 零售平台的完善。因此从大数据视角研究顾客满意不仅对不同类型的企业提高顾 客满意度和竞争力具有实践价值，而且具有数据挖掘意义。 本文基于顾客满意度测评理论，利用结构方程模型，结合顾客评价，构建了 包含品牌形象、顾客期望、网络零售平台、感知质量、感知价值、顾客满意、顾 客抱怨和顾客忠诚的顾客满意度测评模型。针对不同类型的网络零售平台选取京 东、苏宁易购和天猫代表 B2C 平台，淘宝网代表 C2C 平台。利用 python 网络爬 虫采集了苹果、华为、三星、小米和 OPPO 这 5 个品牌共 142 种具体智能数码类 产品的评价数据，其中共有 186656 条有用评价。然后利用基于词典分词、实体 匹配和文本情感分析的方法将采集到的文本数据量化成符合构建的结构方程模 型形式的数值型数据。利用 Amos 将 B2C 和 C2C 平台的数据分别代入初始构建 的结构方程模型，修正并进行模型适配度检验。最终得到了适用于产品评价的 CSI 模型。通过对模型参数的估计以及路径系数的分析，得出不同类型平台顾客 满意度的主要影响因素和二者的异同点。 结合研究结果，利用 4C 理论从顾客需求、沟通、便利性和购买成本四个视 角，针对 B2C 和 C2C 两类网络零售平台分别提出提高顾客满意度的对策。 关键词：顾客满意度，网络零售平台，产品评价，文本情感分析，大数据Abstract Nowadays, online shopping relying on the development of the Internet has become a common social phenomenon. Online retail customers increased year by year, and online retail sales accounted for a growing proportion of the total retail sales of consumer goods in China. The online retail platform is flooded with a large number of product evaluation data, which directly reflect customer satisfaction. In recent years, the utilization efficiency of data resources affects the development of enterprises. Customer satisfaction determines the future development of the online retail market and the improvement of the online retail platform. Therefore, the study of customer satisfaction from the perspective of big data has not only practical value for different types of enterprises to improve customer satisfaction and competitiveness, but also data mining significance. Based on the theory of customer satisfaction measurement, this paper constructs a customer satisfaction measurement model which includes brand image, customer expectation, online retail platform, perceived quality, perceived value, customer satisfaction, customer complaint and customer loyalty by using structural equation model and customer evaluation. According to different types of e-commerce platforms, JD, suning tesco and Tmall are selected to represent B2C, while taobao represents C2C. The python web crawler was used to collect the evaluation data of 142 specific intelligent digital products of five brands including apple, huawei, samsung, xiaomi and OPPO, of which 186,656 were useful. Then, the text data is quantified into numerical data in accordance with the structural equation model based on dictionary segmentation, entity matching and text emotion analysis. Amos was used to substitute the data of B2C and C2C platforms into the initially constructed structural equation model, and the model suitability test was carried out. Finally, a CSI model suitable for product evaluation is obtained. Based on the estimation of model parameters and the analysis of path coefficients, the main influencing factors of customer satisfaction of different types of platforms and the similarities and differences between them are obtainedbined with the research results, from the four perspectives of customer demand, communication, convenience and purchase cost, the 4C theory is used to propose countermeasures to improve customer satisfaction for B2C and C2C online retail platforms. Keywords: customer satisfaction, online retail platform, product reviews, text emotion analysis, big data目 录 1 绪论 ............................................................................................................................1 1.1 研究背景 .............................................................................................................1 1.2 研究目的和意义 .................................................................................................3 1.2.1 研究目的 .......................................................................................................3 1.2.2 研究意义 .......................................................................................................3 1.3 国内外研究现状 .................................................................................................3 1.3.1 顾客满意度模型的相关研究 .......................................................................3 1.3.2 顾客满意度模型测评的相关研究 ...............................................................8 1.3.3 非结构化文本挖掘的相关研究 .................................................................10 1.4 研究的主要内容 ...............................................................................................12 1.5 研究的方法与思路 ...........................................................................................14 1.5.1 研究方法与工具 .........................................................................................14 1.5.2 研究思路 .....................................................................................................14 2 相关研究理论和方法 ..............................................................................................16 2.1 相关研究理论 ...................................................................................................16 2.1.1 基本概念 .....................................................................................................16 2.1.2 顾客满意度指数测评理论 ..........................................................................16 2.1.3 4Cs 营销理论 .............................................................................................. 17 2.2 相关研究方法 ...................................................................................................17 2.2.1 结构方程模型 .............................................................................................17 2.2.2 文本情感分析 .............................................................................................18 2.2.3 基于词典的方法 .........................................................................................19 2.2.4 结巴分词的原理 .........................................................................................19 3 产品评价数据采集及 CSI 测评模型构建 ..............................................................21 3.1 产品评价数据采集 ...........................................................................................21 3.1.1 网络零售平台选择 .....................................................................................21 3.1.2 智能数码类产品的选择 .............................................................................21 3.1.3 网络爬虫分析和数据采集 .........................................................................22 3.2 数据清洗原则及过程 .......................................................................................24 3.2.1 侦辨思路及原则 ............................................................................................24 3.2.2 数据清洗 ........................................................................................................25 3.3 数据分析 ...........................................................................................................25 3.3.1 顾客购买店铺占比分析 ..............................................................................25 3.3.2 顾客购买地区分布 ......................................................................................26 3.3.3 顾客购买途径分析 ......................................................................................27 3.3.4 顾客会员等级 ............................................