Comparative Analysis of Association Rule Mining Techniques for Monitoring Behavioural Patterns of Customers in a Grocery Store
No Thumbnail Available
Date
2015
Journal Title
Journal ISSN
Volume Title
Publisher
African Journal of Computing & ICTs (AJOCICT)
Abstract
The amount of data being generated and stored is growing exponentially, due to the continuing advances in computer technology.
This presents tremendous opportunities for those who can unlock the information embedded within this data, but also introduces
new challenges. The primary challenge is how to discover the hidden knowledge or pattern from the large sets of data in order to
be able to make intelligent decisions that would shape the future of the store and also to determine the best tool to use in mining.
This study proposed to use two different algorithms in association rule mining. The main motivation for carrying out this study is
to compare the two algorithms and determine the time complexity of the two algorithms in mining association rules. In this paper,
association rule techniques were compared and analyzed for monitoring behavioural pattern of customers in a grocery store. The
FP-growth Algorithm and Apriori Algorithm were applied on sixty three (63) datasets from a grocery store. The time complexity
of the two algorithms were considered and it was observed that FP-growth Algorithm is more efficient within the resource
constraints than Apriori Algorithm. The comparison and analysis were implemented using Weka tool. The results revealed that
FP-growth Algorithm is currently one of the fastest approaches for frequent data item set mining.
Description
Keywords
Time Complexity, Fp-growth, Apriori, Association rule mining, Grocery store