This repository contains a Fraud Detection project that utilizes Deep Learning and Self-Organizing Maps (SOM) Neural Networks to identify fraudulent activities in financial transactions. The project focuses on data preprocessing, visualization, and model training using advanced machine learning techniques.
The dataset used in this project is available on my Kaggle page:
π Credit Cards Applications.
- Deep Learning Approach: Implements SOM, an unsupervised neural network for anomaly detection.
- Data Processing: Uses Pandas and NumPy for handling transaction data.
- Visualization: Leverages Matplotlib, Seaborn, and Plotly for data insights.
- Feature Scaling & Clustering: Prepares data for effective fraud detection.