Income Level Prediction

A complete data mining pipeline built to predict income level from an individual's demographic and personal attributes. The project walks the full data science lifecycle: exploratory data analysis and visualization, missing-value and outlier handling, categorical encoding (label + ordinal), feature normalization, class-imbalance correction with SMOTE, and dimensionality reduction through PCA and feature selection. Several classical classification algorithms were trained and compared against a custom artificial neural network built with Keras/TensorFlow, with hyperparameters fine-tuned to identify the best-performing model. Results were interpreted and visualized to surface which demographic factors matter most for income prediction.