Fuzzy Clustering Based Association Rule Mining: A Case Study on Ecommerce

Abstract

Association rule mining (ARM) refers to a procedure that focuses on finding frequent patterns in various data sources. The most commonly used area is the retail sales data and rules which show an association between sales of two products are investigated. To this end, sales data is preprocessed, and algorithms are used to find the association rules by using the specific threshold values of Support and Confidence parameters. In this study, we investigated the effects of using fuzzy clustering with association rule mining A case study from the E-commerce area is selected and sales data for a specific period is analyzed. First ARM is applied to the whole data, then the data is segmented by using Fuzzy Clustering, and ARM is applied to all segments. Later the resulting rules are compared and the effects of segmentation on ARM results are analyzed.