Informal worker classification using classification and regression tree (CART) and non-linear support vector machine (SVM)

Abstract

The problem of decent work availability is still a big challenge in Indonesia. A key element of decent work is employment opportunities where people work in informal sectors are the mains indicator. However, the informal sectors will derive new problems in a long period because their workers are very vulnerable to economic shocks. They do not have a fixed income, lack basic social or legal protections such as health insurance or accident insurance, and no pension plan. Various factors influenced who works in informal sectors. In this study, we want to classify the type of workers working in informal and formal sectors. We used a classification technique called Classification and Regression Tree (CART) and non-linear Support Vector Machine (SVM) to answer such a problem. We implemented both techniques to worker status in the Blitar regency in 2019. We have that the performance of SVM is better than CART due to the classification accuracy score where the CART score is 79.56 percent while the SVM reaches 85.32 percent.