NEW METHOD TO MINING ASSOCIATION RULES USING MULTI-LAYER MATRIX QUADRANT

Abstract

Successful retail organizations utilizing any information they had for managing sales strategies. Most of information about consumer’s retail organization had been stored in transactions database. Discovering knowledge from information stored in the transaction database has led several established methods implemented in many cases with their advantages and disadvantages. One of methodologies to uncover relationship among frequent items purchased in transaction database known as association rules. However, the research of association rules techniques to find knowledge from transaction database still provides a significant opportunity for new methods to participate. In this paper, we proposed a new method of mapping a frequent item set to a multi-layer matrix quadrant. This new method could show the metrics usually used to describe the association rules between items purchased same as any method used in association rules analysis.

Keywords: Association Rules, Matrix Quadrant, Support, Confidence, Lift Ratio