Credit Card Fraud Detection

Design a model to recognize fraudulent credit card transactions.

Dataset

The dataset is available on Kaggle.

• Instead of actual features, PCA components are available in the dataset along with transaction amount and elapsed time.
• Data-Splits: Train/Dev/Test

Analysis

• Perform basic data exploration.
• Check for missing values.
• Check for outliers.
• Perform univariate analysis (using t-test).
• Perform bivariate analysis.

Utilities

• Feature Scaling: RobustScaler is used in logistic regression modeling.
• Feature Selection: Recursive Feature Elimination algorithm is used to determine the best feature combination.
• Data Imbalance: SMOTEENN Hybrid and ClusterCentroid downsampling algorithms have been experimented with to address class imbalance problems.
• Hyperparameter Tuning: GridSearchCV algorithm is used to set the hyperparameter values.

Evaluation

• Area Under Precision-Recall curve is used as the main metric in imbalanced data settings.
• Also, the F1-score is reported.

Results

Model	F1-score	AUPR
Logistic Regression	0.76	0.59
Decision Tree	0.67	0.60
Random Forest	0.72	0.63
XGBoost	0.67	0.61