Association rules application to identify customer purchase intention in a real-time marketing communication tool

To make real-time marketing tools for online storefronts, it is necessary to understand intentions of customers who are connecting on the storefronts. One of important customer intention may be whether a customer intends to purchase or not in the current session. In this paper, we propose customer purchase probability prediction method based on clickstream data using association rule generation techniques. Clickstream data is converted to session data, and the session data is used to generated association and disassociation rules using data mining tools. We propose a method to predict customer purchase probabilities based on the confidence values of the generated association rules. The usefulness of the proposed approach is demonstrated using a real internet bookstore clickstream data set.