基于时间序列交易数据的服装电商客户分类研究

Research for Customer Classification of Clothing E-Business Based on Time Series Transaction Data

摘要: 客户分类是企业进行客户关系管理的前提。文章在充分利用时间序列交易数据的基础上，结合电子商务环境下客户价值研究，建立了时间序列数据与静态数据相结合的多指标分类模型，并提出对时间序列数据进行特征提取处理，应用ReliefF算法对指标选择，最终得服装电商客户多指标加权分类模型。通过某著名服装电商的交易数据的分析，得出了有效客户分类，表明了本方法具有较强的客户分类和客户消费特征的解释能力。

Abstract: Customer classification is the premise of CRM. Based on the time sequence of transaction data and study of customer value in e-business environment, this paper proposed a multi index classification model with time series data and static data. In model application, this paper using feature ex-traction method for the time series data transform, ReliefF algorithm for index selection, then ob-tain a multi index weighted classification model. The results obtained from the real clothing busi-ness show that the customer groups formed using the method clustering all has statistical signifi-cant differences, and with meaningful explanations in terms of marketing strategy. Thus, this study considers useful for discriminative customer relationship management.