ANALISIS CALON PENERIMA BANTUAN PEMERINTAH BAGI PELAKU USAHA MIKRO YANG TERDAMPAK COVID-19 DI KOTA TANJUNG PINANG MENGGUNAKAN METODE CLASSIFICATION AND REGRESSION TREE (CART)

Abstract

This study aims to analyze the factors influencing the eligibility of recipients of the Micro Business Productive Assistance (BPUM) program in Tanjung Pinang City using the Classification and Regression Tree (CART) method. The CART technique was selected due to its ability to identify complex and non-linear relationships among variables without requiring specific statistical assumptions. The research data were obtained from surveys and documentation of micro-business actors, both recipients and non-recipients of BPUM. The variables used include Business Growth (Turnover Growth), Turnover Before COVID-19, Turnover After COVID-19, and several supporting business indicators. The analysis results indicate that the combination of Turnover Growth, Post-COVID-19 Turnover, and Pre-COVID-19 Turnover forms the strongest pattern in determining BPUM eligibility status. The resulting model achieved an accuracy rate of 100 percent, demonstrating an excellent predictive capability. These findings highlight that business performance before and after the pandemic serves as a key benchmark for the government to assess the pandemic’s impact on the microenterprise sector. Moreover, the study confirms that the CART method is effective as a data-driven decision-support tool in improving the accuracy and transparency of government aid distribution. The implications of this study emphasize the importance of integrating statistical and machine learning approaches into public policy design, particularly to enhance objectivity, accountability, and targeting accuracy in micro-business assistance programs.