First Faculty Advisor
Machine Learning; Income
The use of machine learning models to improve prediction problems and handle increasingly large datasets is a rising trend in economics. Prediction plays a particularly important role in applied economics because it provides critical insights to assess market outcomes. This study builds on previous literature to showcase the relative power of these modelling methodologies in economics through the prediction of income. This research utilizes data from the Current Population Survey from 2017 – 2020, containing 467,811 observations and 264 variables. 2017-2018 data served as training data for the models and 2019-2020 served as data for the two testing sets. The results show that machine learning models performed better than traditional prediction approaches in predicting individual total income. The high performance of the machine learning models supports that these methodologies should be utilized alongside more traditional techniques to assist in economic research focusing on prediction. With further development, these models could be used with great effect to assist in both the public and private sectors.