<![CDATA[Transaction data in cash register systems has the potential to serve not only as sales records but also to identify customer purchasing patterns that can support business decision-making. Unfortunately, the utilization of transaction data in coffee shops is often limited to sales reports, resulting in suboptimal use of the relationships between menu items. This study aims to analyze menu purchasing patterns using the Frequent Pattern Growth (FP-Growth) algorithm and to interpret the resulting association rules as a basis for menu recommendations in the cash register system. Employing a quantitative approach and descriptive methods, this research analyzes the public dataset The Bread Basket, which contains over 9,000 transactions. Data preprocessing was conducted using RapidMiner Studio through attribute selection and transformation stages. The FP-Growth algorithm was applied with a minimum support parameter of 5% and a minimum confidence of 20%, yielding two association rules: Bread → Coffee and Cake → Coffee. The Cake → Coffee rule demonstrates a positive relationship with a support value of 0.055 and confidence of 0.527, while Bread → Coffee does not indicate a positive association despite having a higher support value. These findings suggest that the FP-Growth algorithm is effective in identifying purchasing patterns that support menu recommendation logic.]]>
Copyrights © 2026
