Credit Card Fraud detection (Anomaly detection - Imbalanced dataset) - Kaggle

Context

It is important that credit card companies are able to recognize fraudulent credit card transactions so that customers are not charged for items that they did not purchase.

Objectives

Primary: Classify transactions as either spam or non-spam based on 30 input variables (28 of them have been created from a dimensionality reduction exercise)
Secondary: Focus on trying to improve Logistic Regression model through several technics including Grid Search optimizing for F1 and AUPRC

Content

• EDA
• Logistic Regression
  • GridSearchCV
  • Confusion Matrix
  • Classification Report
  • Precision-Recall curve and optimized threshold

For more details see GitHub ML projects - Credit Card Fraud detection - Gregoire LAURENT.