关键词: 企业网络舆情, 客户满意度, 情感分析, 尺度变换, 变尺度聚类

Abstract: With the rapid development of social network platforms, public opinion has become a significant information source for enterprises to excavate valuable information and expand their competitive advantages. This paper studies the problem of enterprise customer relationship management under online public opinion. By constructing the information feedback model of enterprise customers, this paper describes the relationship among enterprise customer, social network user and enterprise public opinion. An algorithm of variable scale clustering based on the information feedback model is also proposed. The algorithm expands the solution of traditional clustering methods via transforming single scale analysis to multi-scale analysis, which overcomes the inconspicuous characteristics problem of clustering results in actual data application. This paper selects Weibo as data source, and implements data analysis experiments through an enterprise network public opinion data set and an enterprise customer data set. The experimental results show that enterprises can achieve customer satisfaction prediction by acquiring public opinion related to their main business. Also, the results of the variable-scale clustering algorithm can provide decision support for enterprises to further develop sales strategies and sales tactics.

Key words: enterprise network public opinion, customer satisfaction, sentiment analysis, scale transformation, variable-scale clustering