Credit Card Fraud Detection — Independent Project (AI model training & evaluation)
Built and evaluated credit-card fraud detection models using supervised classification workflows in Python and SQL. Focused on EDA, cleaning, transformation, and model evaluation using standard metrics and a confusion matrix to support anomaly detection decisions. The work aligns with AI training and labeled-data preparation through systematic data processing rather than manual annotation. • Performed EDA and data cleaning to transform transaction features into model-ready variables • Trained classification models in Scikit-Learn and assessed Precision, Recall, and F1-Score • Used confusion matrix analysis to optimize detection while minimizing false negatives • Designed an end-to-end pipeline for risk mitigation using anomaly detection modeling