大数据视角下中性在线评论对消费者品牌选择的影响研究

发布时间：2018-07-27 17:52

【摘要】：电子商务的蓬勃发展丰富了消费者的购物渠道,网络购物逐渐成为主流的消费方式之一,网络消费者的购物行为吸引了企业家和研究学者共同的关注。传统购物环境下,消费者会依据产品属性和他人推荐做出决策,而网络购物环境下,消费者更多地依赖网络口碑。在线评论是典型的网络口碑,主要包括星级评论和文本评论,过往研究主要从星级评论的角度对在线评论进行分类,探讨不同在线评论对产品销量或者消费者选择的影响,很少有研究从文本评论的角度分析不同在线评论对消费者选择的影响。由于文本评论更容易引起消费者的注意,从文本评论角度入手将在线评论分为正面评论、负面评论和中性评论三类分析其对消费者品牌选择的影响更具研究意义。本研究主要探究了中性在线评论对消费者品牌选择的影响。选取平板电脑这一常见的搜索型商品,利用网页抓取技术,抓取了天猫购物平台多个品牌的平板电脑商品展示网页上的评论数据,清理形成结构化数据。对在线评论进行分词和计算词频处理后,分析得出消费者主要关注的几类平板电脑属性因素,基于既有研究成果,确定了消费者平板电脑品牌选择模型的商品属性变量。借助在线评论的情感强度分析工具,计算文本评论的情感值,选取不同的情感值划分区间,将在线评论分为正面评论、负面评论和中性评论。以特定时段的评论者作为观测个体,按“最畅销”原则选取商品构建选择集,选取消费者个体属性因素和商品属性因素,构建消费者决策树,利用嵌套Logit模型对消费者平板电脑品牌选择进行了分析。实证分析后发现,基于文本评论的情感倾向分析,能有效地将在线评论分为三类,而且中性在线评论对消费者品牌选择存在显著影响,较高比例的中性在线评论会降低消费者的品牌选择概率。另外,消费者个体属性变量和部分商品属性变量对消费者品牌选择也具有显著影响。
[Abstract]:The booming development of electronic commerce enriches the shopping channels of consumers, and online shopping gradually becomes one of the mainstream consumption methods. The shopping behavior of online consumers has attracted the common attention of entrepreneurs and researchers. In the traditional shopping environment, consumers will make decisions based on product attributes and other people's recommendations, while in the online shopping environment, consumers rely more on online word-of-mouth. Online reviews are typical online word-of-mouth, mainly including star reviews and text reviews. Previous studies have classified online reviews from the perspective of star reviews to explore the impact of different online reviews on product sales or consumers' choices. Few studies have analyzed the impact of different online reviews on consumers' choices from the perspective of text reviews. Since text reviews are more likely to attract consumers' attention, online reviews are divided into positive comments from the perspective of text reviews, and negative comments and neutral comments are more meaningful to analyze their impact on consumer brand selection. This study explores the impact of neutral online reviews on consumer brand selection. This paper selects the tablet computer as a common search product, uses the technology of web crawling, grabs the comment data on the display pages of several brands of tablets on the Tmall shopping platform, and cleans up the structured data. After word segmentation and word frequency processing of online comments, several kinds of tablet computer attribute factors that consumers are mainly concerned about are analyzed. Based on the existing research results, the commodity attribute variables of consumer tablet brand selection model are determined. With the help of the affective intensity analysis tool of online comment, the emotion value of text comment is calculated, and different emotional values are selected to divide online comment into positive comment, negative comment and neutral comment. According to the principle of "best selling", the selected items are selected to construct the selection set, and the individual attribute factors and commodity attribute factors of consumers are selected to construct the consumer decision tree. The brand selection of consumer tablet computer is analyzed by nested Logit model. After empirical analysis, it is found that the emotional tendency analysis based on text reviews can effectively classify online reviews into three categories, and neutral online reviews have a significant impact on consumer brand selection. A higher percentage of neutral online reviews will reduce the consumer's brand selection probability. In addition, consumer individual attribute variables and some commodity attribute variables also have a significant impact on consumer brand selection.
【学位授予单位】：哈尔滨工业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：F713.55

【参考文献】