Machine Learning-Based Income Inequality Prediction: A Case Study

Income inequality is an important socioeconomic issue, especially in countries such as the United States, where income distribution gaps continue to widen. Predicting income levels and understanding the factors that contribute to these gaps can provide valuable insights for policymakers and researchers. This paper explores the application of XGBoost (Extreme Gradient Boosting) modeling to income level prediction. The task consists of predicting whether an individual's annual income exceeds $50,000 based on their socioeconomic attributes. After preprocessing the dataset (including dealing with missing values and encoding categorical features), the XGBoost model was trained and optimized using hyper-parameter tuning and cross-validation. The ROC AUC score was 0.93, highlighting the ability of the trained machine learning model to effectively discriminate between income categories.