<![CDATA[This study integrates the Apriori algorithm and local spatial analytics to assess inequality in social assistance distribution in Gorontalo Province, Indonesia, covering six regencies/cities. The administrative Beneficiary Master List (BNBA) dataset was standardized for association rule mining to identify co-beneficiary patterns across major social assistance schemes, namely PKH, BPNT, BST, and BPUM. In parallel, the data were aggregated at the sub-district level to construct an inequality score based on Principal Component Analysis (PCA) of beneficiary proportions, which was then analyzed using Local Moran’s I (LISA) and Getis–Ord Gi*. The Apriori analysis of the province-wide dataset produced 64 association rules, 74 frequent itemsets, and 38 unique items. The results indicate strong co-beneficiary relationships among BPNT, BST, BPUM, and PKH, with confidence values ranging from approximately 0.60 to 0.95 and lift values exceeding 10. Spatial analysis shows that five of the six regencies/cities exhibit significant positive spatial autocorrelation (p < 0.10), with particularly strong clustering in Pohuwato (I = 0.9681) and North Gorontalo (I = 0.8331), while Gorontalo Regency shows no statistically significant pattern. LISA cluster maps further identify high-high (HH) clusters in parts of Boalemo and North Gorontalo, as well as low-low (LL) and high-low (HL) areas relevant for policy refinement. These findings suggest that integrating Apriori and local spatial analytics provides an effective operational approach for improving targeting accuracy, reducing overlap in assistance allocation, and identifying areas at risk of under-coverage.]]>
Copyrights © 2026
