FraudDetection

Overview

FraudDetection is a Python-based system for detecting fraudulent transactions in financial data. The system uses a combination of machine learning algorithms and data preprocessing techniques to identify potentially fraudulent transactions.

Features

• Data Preprocessing: The system includes a data preprocessing module that handles missing values, encodes categorical variables, and normalises numerical features.
• Machine Learning Models: The system includes two machine learning models: a deep learning model (FraudDetectionModel) and a random forest model (FraudDetectionRFModel).
• Model Evaluation: The system includes a model evaluation module that calculates accuracy, precision, recall, and F1 score for each model.
• Model Saving and Loading: The system allows for saving and loading trained models for future use.

Requirements

• Python: The system is built using Python 3.10.14.
• Libraries: The system requires the following libraries:
  • pandas for data manipulation and analysis
  • numpy for numerical computations
  • torch for deep learning
  • sklearn for machine learning
  • matplotlib and seaborn for data visualisation

Usage

Data Preparation

1. Download the dataset (e.g., Fraud.csv) and place it in the Fraud_data directory.
2. Preprocess the data using the DataPreprocessing module.

Model Training

1. Train the deep learning model using the FraudDetectionModel class.
2. Train the random forest model using the FraudDetectionRFModel class.

Model Evaluation

1. Evaluate the performance of each model using the ModelEvaluation module.

Model Saving and Loading

1. Save the trained models using the ModelSaving module.
2. Load the saved models using the ModelLoading module.

Example Use Cases

• Fraud Detection: Use the system to detect fraudulent transactions in a financial dataset.
• Model Comparison: Use the system to compare the performance of different machine learning models on a fraud detection task.

Contributing

Contributions are welcome! If you would like to contribute to the FraudDetection system, please fork the repository and submit a pull request.

License

The FraudDetection system is released under the MIT License.