Determining rural development priorities using a hybrid clustering approach: a case study of South Sulawesi, Indonesia

Abstract

This study aims to develop village clustering techniques that facilitate the government's placement of experts in villages identified as development priorities. The clustering process utilizes indicators such as the community standard of living index (CSLI), head of family, and number of residents. The CSLI was constructed by involving 900 respondents from 30 villages in South Sulawesi Province, Indonesia. The clustering technique is constructed using a hybrid approach of self-organizing map (SOM), Xie-Beni, and fuzzy c-means (FCM) methodologies. The resulting clusters are categorized into three groups: CSLI-poor, which comprises three sub-clusters of CSLI-poor, CSLI-average, and CSLI-excellent. To determine recommendations for the required field of expertise in each village, the cosine similarity algorithm is employed. Villages classified within the CSLI-poor cluster are considered development priorities. The findings revealed that 36.7% of villages were classified as CSLI-poor, 26.7% as CSLI-average, and 36.7% as CSLI-excellent. Consequently, all villages require experts in the fields of economics, social sciences, and health sciences.