Exploring Supervised and Unsupervised Learning Techniques in Market Basket Analysis

Abstract

Market Basket Analysis is an important task in data mining. It is widely used in retail sector to reveal customer buying habits by finding associations among items in transaction data. Traditionally, unsupervised methods like Apriori or FP-Growth have been used to identify frequent itemsets and association rules in Market Basket Analysis. But with the advancement of analytical techniques, supervised learning methods have become more popular for their predictive capabilities. This paper compares supervised and unsupervised data mining methods in Market Basket Analysis. It highlights their applications, strengths, limitations and future possibilities. The paper also discusses hybrid and Fuzzy techniques to provide a wider view of modern market basket analysis.