Detect fraudulent credit card transactions (Class = 1) using machine learning.
The project focuses on accurate identification of fraud in a highly imbalanced dataset.
This project employs a Random Forest Classifier optimized with RandomizedSearchCV.
Key points:
- Handles imbalanced data effectively
- Provides probability scores for fraud prediction
- Threshold adjustment available for recall optimization
- Checked for missing values and duplicates
- Selected
Classas the target column - Split data into training and testing sets
- Visualized distributions and relationships using scatter plots, line plots, KDE, and boxplots
- π Python 3
- π pandas, numpy for data handling
- π matplotlib, seaborn for visualization
- π€ scikit-learn for modeling
- π Jupyter Notebook for interactive analysis
- π GitHub: Mehdipoladrag
- π Kaggle: mehdip1