Employment is one of the problems of every developing country, including Indonesia. Job creation is a very important issue for economic development. The Province of the Special Region of Yogyakarta is a province that is quite successful in overcoming labor problems with the third highest TPAK after Bali and Papua in 2018 with a percentage of TPAK that has tended to be stable since 2015. However, it turns out that the percentage of poor people who work in the informal sector in Yogyakarta is very large, reaching 44.97%. , almost half of the total. The income level of workers in the informal sector is much lower than that of workers in the formal sector. Data on the income level of informal workers was identified as unbalanced data because the comparison between workers with low and non-low incomes was very unequal. Therefore, the SMOTE method is used to overcome the problem of unbalanced data. The classification method used in this study is binary logistic regression. This study aims to determine the variables that affect the income level of informal workers and compare the models before and after SMOTE to get the best classification model. The results of the model evaluation show that the model after SMOTE is better at classifying the income level of informal workers. Furthermore, the variables that affect the income level of informal workers are classification of residence, gender, marital status, job training, education level, business field, age, and working hours..
Copyrights © 2021