How mondrita/credit-card-fraud-detection works
Project type and purpose
• This is a Machine Learning exploration project (not a multi-tier web app or microservice suite).
• Purpose: build and compare several classification models (KNN, Logistic Regression, SVM, Naive Bayes) to detect credit-card fraud on an imbalanced dataset.
File structure
• Top level:
• CREDIT_CARD__FRAUD.ipynb – the single Jupyter notebook containing all steps
• card_transdata.csv – raw transactions dataset
• README.md – project description, objectives, metrics, dependencies
• No separate “src” or “tests” directories, so the code, data loading, preprocessing, modeling, and visualization all live in the one notebook.
README for additional insights
• Workflow described implicitly: data exploration → preprocessing (handling imbalance, scaling, etc.) → model training → evaluation (precision, recall, F1) → comparative analysis.
• Dependencies (pandas, numpy, matplotlib/seaborn, scikit-learn) confirm a standard Python data-science stack.