Implementation of Data Mining Using the FP-Growth Algorithm in Analyzing Sales Transactions at Difour Gabriel Store

Abstract

Difour Gabriel Store, located in Buntao', is a retail store specializing in daily household necessities. Established in January 2022 along the main transportation corridor connecting Buntao' and Rantebua, the store experiences a continuous increase in daily customer traffic and transaction volume. To stay competitive, understanding customer buying patterns and preferences is essential. This study implements data mining using the Frequent Pattern Growth (FP-Growth) algorithm to discover association rules and identify products frequently purchased together. The FP-Growth algorithm constructs an FP-Tree data structure to efficiently extract frequent itemsets from sales transaction data. The results demonstrate that FP-Growth effectively uncovers strong purchasing patterns. Specifically, a significant association rule was found between Qitela Barbeque 20G and Permen Pendekar Biru, achieving a lift ratio of 85.00. Furthermore, evaluations across varied confidence thresholds (0.1, 0.3, 0.5, 0.7, and 0.9) revealed that higher confidence thresholds produce more selective and highly reliable association rules. These findings provide actionable decision-support insights for store management to optimize product shelf placement, cross-merchandising strategies, and targeted promotional campaigns.