Implementation of Multiple Discriminant Analysis (MDA) for Clustering Smart Village in West Java Based Podes (Potensi Desa) Database

Abstract

Village development is accelerated through the smart village program. Village development in Indonesia, including West Java Province, is measured using the Developing Village Index (IDM). The 50 villages that are the locus of smart villages in West Java Province consist of three statuses, namely underdeveloped, developing and advanced. By using clustering using Multiple Dicriminant Analysis (MDA), this study try to identify smart village locus indicators based on the Village Building Index (IDM) based on Village Potential Data and classify them based on their status. The results of the smart village clustering in West Java province show that the smart village program can be adopted by villages with a lagging village index (IDM) level of 12%. Nine predictors were found to distinguish intelligent villages with lagging, developing and independent status, namely Public places, hospitals, pharmacies, community activities, security posts, banks & financial institutions, agricultural stores and Sewerage Treatment. This condition is a very interesting finding because it breaks the assumption that smart villages can only be adopted by villages with a minimally developed IDM level. The results of smart village clustering based on village potential data also show that smart village clusters are dominated by developing IDM-level villages. The multiple discriminant-based clustering model was carried out through two alternative models, and one of the models was able to show a canonical correlation of 0.906.