Apriori-Based Data Mining of Sales Transactions at Buya Aqiqah MSME

Abstract

Buya Aqiqah is a micro, small, and medium enterprise (MSME) providing aqiqah—an Islamic ceremonial catering service traditionally celebrating a child's birth—and general catering in Depok, Indonesia, since 2003. Although the enterprise has accumulated substantial sales transaction data, these records have previously served solely as manual administrative archives rather than an analytical foundation for identifying customer purchasing patterns to support managerial decision-making. This study aimed to apply the Apriori algorithm to discover association patterns among menu items using Buya Aqiqah's historical sales transaction records. Guided by the Knowledge Discovery in Databases (KDD) framework, the analysis was executed using RapidMiner Studio. A total of 137 digitized sales transactions were selected, cleaned, transformed into binary format, and evaluated across three parameter scenarios: Scenario 1 (minimum support 5%, minimum confidence 50%), Scenario 2 (minimum support 10%, minimum confidence 60%), and Scenario 3 (minimum support 15%, minimum confidence 70%). Across all three scenarios, two valid association rules consistently emerged: Gulai → Sate (support 67.2%, confidence 94.8%, lift ratio 1.065) and Sop → Sate (support 19.7%, confidence 100%, lift ratio 1.123). Scenario 1 was determined as the most representative parameter setting because it captured the broadest frequent itemset coverage without omitting any valid rules. The results position Sate as the core menu item consistently paired with other dishes, confirming that the identified associations reflect genuine purchasing behaviors rather than random co-occurrences. These findings provide empirical evidence for management to design targeted bundling packages—specifically Gulai–Sate and Sop–Sate—while prioritizing inventory procurement for core ingredients.