Enhanced MSME Support Allocation with Integrated K-means and Tukey's Outlier Detection

The pandemic caused by the Covid-19 virus has resulted in Indonesia experiencing a slowdown in economic growth. MSMEs (Micro, Small and Medium Enterprises) as the majority contributor to the National GDP and absorb more than 117 million workers have not been spared the negative impacts of the pandemic. This study addresses the challenge of misdirected governmental support to MSMEs (Micro, Small, and Medium Enterprises) in Indonesia during the COVID-19 pandemic, a period marked by significant economic downturns. Despite the government's efforts to bolster MSMEs through various incentives and subsidies, discrepancies in aid distribution were evident, with 414,612 non-targeted beneficiaries receiving funds. To enhance the precision of aid allocation, this research employs a novel approach by integrating the K-means clustering algorithm with Tukey's method for outlier removal. This methodology effectively segments Indonesian MSMEs into three distinct clusters based on annual sales, total assets, and workforce size, thereby facilitating more targeted support strategies. The clusters are prioritized for assistance, with Cluster 3 identified as the highest priority due to its lowest sales figures, followed by Cluster 2 and Cluster 1, based on their respective economic performances. The study's findings, underscored by a Silhouette Coefficient of 0.526, confirm the efficacy of this approach. Furthermore, it suggests specific sectors and regions for focused support, highlighting wholesale, retail trade, and repair services of cars and motorcycles in East Kalimantan as primary beneficiaries. This research contributes to the optimization of MSME support mechanisms during crises, offering a replicable model for targeted assistance based on empirical data analysis.