Discovering Important Rules and Loyal Customer by Considering the Repetition in Association Rules

In this paper, we present a study on association rule mining based on real POS data from a SME (small and medium enterprise). Our main contribution is proposing a new criterion, called R-repeat, to evaluate the significance of rules by focusing attention to the purchase pattern repeated by a certain customer. We also focus on detecting a special customer segment, called R-repeater. In addition, we make a comparison between R-repeater segment and the customer segments by REM analysis, and confirm that R-repeater is a loyal customer segment