Clustering Algorithm for Market-Basket Analysis: The Underlying Concept of Data Mining Technology

Abstract

The goal of data mining is to extract interesting correlated information from large databases. This thesis seeks to understand the underlying concept of data mining technology in market-basket analysis. The clustering algorithm based on Small Large Ratios, SLR is presented in a manner that helps to understand the concept of data mining technology in marketbasket analysis. The author used a data mining software called PolyAnalyst 4.5 to perform analysis on the set of items that customers have bought in supermarket for market-basket application. In this research, the author tried to relate the algorithm presented with the experiment. Then, the author discussed the results by showing an application of marketbasket analysis. The statistical results from the PolyAnalyst's reports are explained and elaborated further i n the results section. The author summarized the findings and tried to relate them to the benefits of data mining towards organizations.