Comparative Analysis of Apriori and FP-Growth for Order Pattern Analysis to Support Sales Strategy at Buya Aqiqah

Intense market competition and suboptimal transaction data management have contributed to declining sales at Buya Aqiqah, a small and medium-sized enterprise (SME) specializing in aqiqah and catering services in Depok, West Java. To address this business challenge, this study conducts a comparative performance and output analysis of the Apriori and Frequent Pattern Growth (FP-Growth) algorithms to extract customer purchasing patterns and formulate data-driven sales strategies. Following the Cross-Industry Standard Process for Data Mining (CRISP-DM) framework, the research processed historical sales transaction records spanning a two-year operational period from April 2024 to April 2026. The dataset was transformed from 749 raw sales entries into 541 tabular transactions comprising 79 distinct menu items. Experimental evaluations conducted in Google Colaboratory using Python across five distinct minimum support and confidence thresholds demonstrated that FP-Growth consistently outperformed Apriori, exhibiting substantially faster execution times and superior memory efficiency due to its compact tree-based representation without repetitive candidate itemset generation. Under the optimal parameter baseline of 2% minimum support and 100% minimum confidence, both algorithms yielded an identical set of 22 valid association rules, differing only in the sequence of rule extraction. The extracted rules identified strong affinity relationships among specific staple and side-dish combinations with significant lift values, particularly involving goat dishes, rice, and complementary menu items. These empirical findings provide actionable strategic guidance for Buya Aqiqah to optimize commercial performance through structured product bundling, cross-selling, up-selling, and targeted promotional discounts tailored to actual customer dining preferences.