<![CDATA[AbstractAccurate village profile planning needs to be done with mapping based on its characteristics. Village grouping based on the characteristics of village facilities and potential or its characteristics based on the Building Village Index indicator can help determine priorities in village development. In this study, mixed data was used, with numerical data grouping using Hierarchical Agglomerative Nesting (AGNES) algorithm and categorical data with Quick Robust Clustering Using Links (QROCK). The resulting clusters are then combined using the QROCK Ensemble algorithm (algCEBMDC). The data is sourced from the 2021 Village Potential Data Collection (PODES) by the Central Statistics Agency in 277 villages in Banjar Regency, including 18 numerical variables and 29 categorical variables. The results of the study obtained optimal clusters based on the ratio of within-group standard deviation (SW) to between-group standard deviation (SB) resulting in a ratio of 4.82.10-9 with a threshold of 0.4 to 0.9 resulting in 6 clusters. The best cluster results are cluster 4 (4 villages) and cluster 3 (14 villages), then cluster 2 (villages) and cluster 1 (186 villages), and clusters that need development priorities are cluser 5 (2 villages) and cluster 6 (1 village) which are outliers based on processing results. Keywords: Village Grouping, Cluster, AGNES Algorithm, Quick Robust Clustering Using Links (QROCK), AlgCEBMDC]]>
Copyrights © 2024
