Anomaly Detection in Credit Card Transactions

This project implements anomaly detection in credit card transactions using the Isolation Forest algorithm. The dataset used includes transaction details such as year, month, department, division, merchant, and transaction amount.

1. Installation
2. Usage
3. Code Structure
4. Features
5. Contributing
6. License

Installation

1. Clone the repository:

   git clone https://github.com/your_username/your_repository.git
   cd CreditCard-Transaction-Anomaly-Detection-Using-IsolationForest

2. Install the required libraries using the environment.yml file using conda:

   conda env create -f environment.yml

Usage

1. Load the dataset:

   Get the dataset from kaggle using https://www.kaggle.com/datasets/ybifoundation/credit-card-transaction Ensure the dataset CreditCardTransaction.csv is placed in the appropriate directory or update the file path in the code (pd.read_csv("drive/MyDrive/CreditCardTransaction.csv")).

2. Run the notebooks:

   Run the notebooks using the conda environment.

3. Explore the results:

   • The code performs exploratory data analysis, preprocessing, applies Isolation Forest for anomaly detection, and visualizes anomalies detected.

Code Structure

• creditcard_transaction_anomaly_detection.ipynb: Main script that loads processed data, applies Isolation Forest, and visualizes anomalies.
• data_exploration.ipynb: notebook that explores the data.
• data_preprocessing.ipynb: notebooks that prepares data for training.

Features

• Data Loading: Loads credit card transaction data from a CSV file.

• Exploratory Data Analysis: Analyzes data distribution, checks for missing values, and explores relationships between variables.

• Preprocessing: Handles missing values, encodes categorical variables using LabelEncoder, and normalizes numerical features using z-score method.

• Isolation Forest: Utilizes Isolation Forest algorithm for unsupervised anomaly detection.

• Visualization: Uses matplotlib and seaborn for visualizing data distributions, scatter plots, and anomalies detected.

Contributing

Contributions are welcome. For major changes, please open an issue first to discuss what you would like to change.

License

This project is licensed under the Apache-2.0 License - see the LICENSE file for details.

Acknowledgments

• Inspired by tutorials and examples on anomaly detection and Isolation Forest.