Analisis Pola Menu Program Makan Bergizi Gratis Menggunakan Algoritma Apriori

Abstract

The Free Nutritious Meals Program (MBG) requires data-driven evaluation to identify the regularity, diversity, and co-occurrence of menu components. This study aims to explore associations among MBG menu items and beneficiary groups using association rule mining with the Apriori algorithm. The research follows the Knowledge Discovery in Databases framework, covering data selection, preprocessing, binary transaction transformation, frequent-itemset mining, rule generation, evaluation, and visualization. The available dataset is a secondary illustrative transaction recap consisting of five menu-distribution records. A minimum support of 40% and a minimum confidence of 60% were applied; rule quality was assessed using support, confidence, and lift. Apriori was selected rather than FP-Growth or Eclat because the dataset is small, the item vocabulary is limited, and the level-wise candidate process is easier to audit and interpret for exploratory policy evaluation. The analysis produced 31 frequent itemsets and 67 association rules before redundancy filtering. Representative positive associations included Fish → Fruit, Milk → Elementary School, and Elementary School → Green Vegetables, while Rice → Green Vegetables had the highest support but a neutral lift of 1.00. Scatter and network visualizations clarified the strength and structure of the rules. The results demonstrate the usefulness of Apriori for transparent exploratory analysis, but the small illustrative dataset means that the findings should not be generalized to national MBG implementation. Larger multi-period datasets and direct comparisons with FP-Growth and Eclat are recommended.