Machine Learning
Credit Card Fraud Detection
Imbalanced classification with metrics.
Project overview
Credit Card Fraud Detection is a mentor-guided ML project designed for university submission. ML pipeline for fraud detection with SMOTE-style balancing notes, ROC/AUC evaluation, and a Flask prediction API. The build uses Python, scikit-learn, Pandas, Flask and is structured so you can explain architecture, modules, and test cases during viva. Typical delivery timeline is 6–8 weeks depending on package and college deadlines.
Best suited for: Students working on predictive analytics and applied ML projects
Tech stack
Objectives
- Design and implement a working solution for “Credit Card Fraud Detection” aligned to academic evaluation criteria.
- Demonstrate core features such as Train pipeline, Metrics, Predict API.
- Produce documentation (synopsis, report chapters, and PPT) suitable for guide approval.
- Prepare viva-ready explanations for design choices, tech stack, and future enhancements.
Key modules
- Module: Train pipeline
- Module: Metrics
- Module: Predict API
- Documentation, deployment notes, and viva Q&A
Key features
- Train pipeline
- Metrics
- Predict API
- Sample data
Deliverables
- Notebook
- Model
- API
- Report
- PPT
Learning outcomes
- Runnable source code with setup guide
- University-style report draft
- Presentation deck (PPT)
- Mentor review checkpoints
