Implementation of Association Rule Mining Using the FP-Growth Algorithm on Non-Procedural Indonesian Migrant Worker (PMI) Data in South Sulawesi

Abstract

This study examines patterns in non-procedural Indonesian Migrant Workers (PMI) data in South Sulawesi and utilizes them to strengthen data-driven prevention. The method used is association rule mining using the FP-Growth algorithm within the CRISP-DM framework implemented through Altair AI Studio software. Modeling is run based on a minimum support value of 30% with a minimum confidence of 80%) for the Makassar, Pare-Pare, and Palopo Immigration Office datasets. Patterns are retained if the lift value is > 1 and selecting the top 10 patterns for each dataset. The results show consistent frequent itemsets and association rules indicating a general pattern of non-procedural PMI dominated by adult males with destinations in Malaysia with illegal/undocumented issues. The findings can be used as a preventive measure in strengthening interviews and document verification in the passport issuance process and the Immigration fostered village program. The study confirms that the application of FP-Growth with support, confidence, and lift evaluations provides evidence-based insights relevant to a more targeted and effective non-procedural PMI prevention policy by the Immigration Office