基于TF-IDF算法和K-means聚类的商品评论与价格波动相关性研究——以ThinkPad电脑为例  

Research on the Correlation of Products Reviews and Price Fluctuation based on TF-IDF and K-means Clustering-Taking ThinkPad as an Example

作　　者：刘家成 王艺憬 孙燕红[1] LIU Jiacheng;WANG Yijing;SUN Yanhong(SHU-UTS SILC Business School,Shanghai University,Shanghai 201899,China)

机构地区：[1]上海大学悉尼工商学院,上海201899

出　　处：《科技创业月刊》2018年第7期45-49,共5页Journal of Entrepreneurship in Science & Technology

摘　　要：随着电子商务产业的发展和电商平台的涌现,消费者逐渐习惯了通过电商平台比对货物和购买商品。相关产品主页上的评论和评分成为了消费者决策的重要参照,同时商家也把评论和评分作为调整价格的重要依据。以ThinkPad E570c电脑为例,运用TF-IDF算法、Kmeans聚类和SPSS统计分析,发现在仅考虑评论属性的前提下,评论中的带图数量与价格波动具有关联性。在同时考虑评论属性和内容的情况下,评论中关性能的内容是消费者的主要关注点。研究结果对商家制定定价策略与合理管理在线评论具有一定的实用意义。The development of e-commerce industry leads the emergence of e-commerce platform. As a result, consumers have become accustomed to comparing and purchasing goods on e-commerce platforms. The relative comments and rates of products on the homepages become vital reference when consumers make decision. Meanwhile, comments and scorings of products are also the most important reference points when the merchants revise the price. This paper takes ThinkPad E570c as an example and analyzes the comments by applying TF-IDF、K-means Clustering and SPSS statistical analysis. The price of product is related to the pictures contained in the comments when only the comment attributes are being considered. The main concern of consumers is the computer performance related contents in the reviews when all the comment attributes and contents are being taken into consideration. The implication and practicability of this study lies in its providing the rational for the sellers to set the pricing strategies and to establish effective ways to manage the on-line comments.

关 键 词：在线评论 文本内容倾向性 TF-IDF 

分 类 号：F713.55[经济管理—市场营销] F724.6[经济管理—产业经济]