Abstract
This research attempts to assist a very large retailer in the midwest of the United States in understanding its consumers’ buying patterns and behaviors by employing a multi-technique data mining approach. Three data mining techniques, including cluster analysis, predictive (neural networks) modeling, and market basket analysis methods, are reviewed for the purpose of comparing the prevalence of their use and effects. This applied multi-technique approach provides a comparative analysis between sales of infant items at the baby centers and non-baby centers, helps understand the buying patterns at these centers, and analyzes items that were most frequently purchased together by consumers. This paper concludes with the research results and recommendations as well as directions for future research.