K-MEANS WITH PARTICLE SWARM OPTIMIZATION FOR ERROR REDUCTION IN MICRO, SMALL, MEDIUM ENTERPISE CRAFT IN YOGYAKARTA

Abstract

Purpose: Micro, Small, and Medium Enterprises (MSMEs) in the handicraft sector of Yogyakarta face significant 
challenges regarding capital access and marketing optimization. This study aims to group 146 MSMEs based on 
business characteristics to support the formulation of targeted empowerment strategies. 
Methods/Study design/approach: A quantitative approach was employed using survey data processed through 
encoding and normalization. The research utilized the K-Means algorithm, optimized with Particle Swarm 
Optimization (PSO) to determine the optimal number of clusters, and evaluated the model using Sum of Squared Error 
(SSE) and Mean Absolute Error (MAE). 
Result/Findings: The results show that the K-Means method optimized with PSO significantly outperforms the 
standard K-Means algorithm. Specifically, the optimized model achieved an SSE of 51.676 and an MAE of 0.116, 
compared to the standard K-Means algorithm which produced a higher SSE of 54.555 and an MAE of 0.124. 
Novelty/Originality/Value: The novelty of this study lies in the application of PSO to minimize clustering errors 
specifically within the Yogyakarta handicraft sector context. These findings offer a highly accurate, data-driven 
foundation for policymakers to design effective MSME development programs.