An approach for K-modes cluster analysis to examine direct resource transfers in brazilian public agreements for innovation and rural development

This article presents an innovative approach to K-modes cluster analysis applied to public agreements signed by the Secretariat of Innovation, Rural Development, and Irrigation (SDI) of the Ministry of Agriculture, Livestock, and Supply (MAPA) of Brazil. The agreements were signed between 2019 and early 2023. The main goal is to identify patterns and trends in resource transfers for innovation and sustainable rural development in Brazil. This study analyzed 10,098 agreements using the Design Science Research Methodology, refined over three cycles. The K-modes method is particularly suitable for handling categorical data. It enables the clustering of agreements based on similar characteristics, such as geographic region served, purpose, the year the agreement was signed, the total value of expense items, and the status of the agreement. The results demonstrate that the K-modes approach overcomes typical limitations of traditional clustering methods, including sensitivity to outliers, restriction to numerical data, and difficulty handling clusters of varying sizes and densities. Additionally, it addresses issues related to the lack of interpretability of the generated clusters. This study advances the application of the K-modes method in analyzing direct resource transfer mechanisms for innovation and rural development, a still unexplored area. The proposed approach can be generalized to examine direct resource transfer mechanisms in different countries, contributing to enhancing public policies on a global scale. Despite structural differences, resource transfers aimed at innovation and rural development aim to direct public funds to initiatives that improve living conditions and the agricultural sector’s competitiveness.