UNCOVERING SHOPPING PATTERNS IN BIG DATA ENVIRONMENTS
Authors
Keywords
market basket analysis, association and sequential rule mining, sales transactions, big data analytics, shopper behavior analysis
Summary
In this study, we present the feasibility of applying selected algorithms to rapidly detect regular, common, and recurring patterns of shopping behavior for products or services purchased simultaneously and/or in some temporal sequence. Such analyses are often defined by the term „'market basket analysis“. The object of the present study is the analysis of large data sets of sales transactions observed across multiple individual acts of purchase, and the subject matter is the discovery of the capabilities of some numerical algorithms from the field of machine learning to detect hidden patterns in acts of purchase by processing data from individual customer „market baskets“. The goal is to convince the reader of the possibilities of extracting association rules from big data through demonstrations with publicly available open data. The exposition follows a „concept-to-application“ logic and consistently provides answers to the questions of „why“ it is necessary to do it, „what“ it is done with, and „how“ it is done. After a brief introduction to the logic and specifics of the most popular algorithms for discovering patterns by extracting association rules from large datasets, we provide detailed working procedures and instructions for analyzing and interpreting the analytical results. Finally, we provide a synopsis and provide guidelines and recommendations for using the analytical procedures in a marketing context.
Pages: 1
Price: 2 Points