Credit Card Fraud Detection Project

This project aims to build a machine learning model to detect fraudulent credit card transactions. It uses a Random Forest classifier trained on a dataset of credit card transactions.

Project Structure

Directory structure:
    ├── README.md
    ├── requirements.txt
    ├── models/
    │   └── random_forest_model.pkl
    ├── notebooks/
    │   ├── 01_data_preprocessing.ipynb
    │   └── 02_model_training.ipynb
    └── outputs/
        └── evaluation_report.txt

Workflow

1. Data Preprocessing (notebooks/01_data_preprocessing.ipynb):

   • Loads the raw dataset (data/creditcard.csv).
   • Performs basic data exploration (info, description, null checks).
   • Visualizes class distribution and correlations.
   • Applies StandardScaler to 'Amount' and 'Time' features.
   • Drops original 'Time' and 'Amount' columns.
   • Saves the processed data to data/processed_data.csv.

2. Model Training (notebooks/02_model_training.ipynb):

   • Loads the processed data (data/processed_data.csv).
   • Splits the data into training and testing sets.
   • Trains a RandomForestClassifier model.
   • Evaluates the model using classification report, confusion matrix, and ROC AUC score.
   • Saves the trained model to models/random_forest_model.pkl.
   • Saves the evaluation results to outputs/evaluation_report.txt.
   • Generates and saves the ROC curve plot to outputs/roc_curve.png.

Setup and Installation

1. Clone the repository:

   git clone <your-repository-url>
   cd <repository-directory>

2. Create a virtual environment (recommended):

• Linux/MacOS

 python -m venv venv
 source venv/bin/activate

• Windows

 python -m venv venv
 venv\Scripts\activate

4. Install dependencies:
   • Ensure the requirements.txt file lists all necessary packages. Based on the notebooks, you'll likely need: pandas, numpy, matplotlib, seaborn, scikit-learn.
   • Install using pip:

     pip install -r requirements.txt

5. Add Data:
   • Place the raw dataset file (creditcard.csv) into the data/ directory. (Note: This dataset is often found on Kaggle).

How to Run

1. Ensure the creditcard.csv file is in the data/ directory.
2. Run the data preprocessing notebook:
   • Execute the cells in notebooks/01_data_preprocessing.ipynb. This will generate data/processed_data.csv.
3. Run the model training notebook:
   • Execute the cells in notebooks/02_model_training.ipynb. This will train the model, save it to models/, and generate evaluation files in outputs/.

Results

The model performance metrics (precision, recall, F1-score, confusion matrix, AUC score) are available in outputs/evaluation_report.txt. The ROC curve visualization is saved as outputs/roc_curve.png