Overview

This project uses BERT (Bidirectional Encoder Representations from Transformers) to classify messages as spam or not spam. The primary goal is to provide a robust and accurate model for spam detection using state-of-the-art natural language processing techniques.

Features

• Leverages BERT for advanced text understanding and classification
• Developed and trained in Jupyter Notebook for easy experimentation and iteration
• Clean, well-organized, and reproducible codebase
• Easily customizable for different spam datasets

Requirements

• Python 3.7+
• Jupyter Notebook
• PyTorch or TensorFlow (depending on your chosen BERT implementation)
• Transformers library (pip install transformers)
• pandas, numpy, scikit-learn

Getting Started

1. Clone the repository:

   git clone https://github.com/krakos-afk/SPAM-classifier.git
   cd SPAM-classifier

2. Install dependencies:

   pip install -r requirements.txt

3. Open Jupyter Notebook:

   jupyter notebook

   Open the main notebook and follow the instructions to load your data, train the model, and evaluate its performance.

Usage

1. Prepare your dataset (CSV or similar format) with labeled spam/not-spam messages.
2. Update the notebook with the path to your dataset.
3. Run the notebook cells to train the classifier.
4. Use the trained model to classify new messages.

License

This project is licensed under the MIT License.