A novel approach to dynamic profiling of e-customers considering click stream data and online reviews

Abstrak

In this paper, we present an approach for mining change in customer’s behavior for the purpose of maintaining robust profiling model over time. Most of previous studies leave important questions unanswered: In developing B2C e-commerce strategies, how do managers implicitly load customer’s profiles based on their satisfaction over the online store characteristics? And: What kind of feedback segments do they have? Our proposed approach does not force customers to explicitly express their preference information over the online service but rather capture their preference from their online activities. The challenge does not only lay in analyzing how customer’s classifier model change and when it does so but also to adapt it to the customer’s click stream data using a new decision tree generation algorithm which takes as inputs new set of variables; categorical, continuous and fuzzy variables. Customer’s online reviews rates are considered as classes. Experiments show that this work performed well in identifying relevant customer’s stream data to judge the chinese e-commerce website “Tmall”. The extracted values of the website’s features are also useful to identifying the satisfaction level when the customer’s rate is not available.