EEG_GPT_Model

This repository contains all data and code used for the paper titled: "ChatGPT-Based Model for Controlling Active Assistive Devices Using Non-Invasive EEG Signals"

Project Structure

EEG_GPT_Model/
├── data/
│   ├── raw/           # Raw EEG and MoCap data
│   └── processed/     # Processed and synchronized datasets
├── results/           # Model evaluation results and visualizations
└── src/
    ├── data_processing/  # Data parsing and synchronization
    └── models/          # Model training and prediction

Data Setup

Before running any code, you need to extract the data files:

1. Extract the raw data:

   # Extract EEG data
   unzip "data/raw/eeg/RAW EEG DATA FINAL.zip" -d data/raw/eeg/
   
   # Extract MoCap data
   unzip "data/raw/mocap/RAW MOCAP DATA FINAL.zip" -d data/raw/mocap/

Setup

1. Create a virtual environment (recommended):

python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate

2. Install dependencies:

For data processing:

pip install -r src/data_processing/requirements.txt

For model training and prediction:

pip install -r src/models/requirements.txt

Usage

1. Data Processing:

   • Synchronize datasets: python src/data_processing/synchronize.py --mocap_dir data/raw/mocap --eeg_dir data/raw/eeg --output_dir data/processed

2. Model Training:

   • Train the model: python src/models/train.py

3. Model Prediction:

   • Run predictions: python src/models/predict.py

Data

• Raw EEG data is stored in data/raw/eeg/
• Raw MoCap data is stored in data/raw/mocap/
• Processed and synchronized datasets are in data/processed/

Results

The results/ directory contains all model evaluation outputs:

• Training and validation plots
• Direction classifier visualizations
• Joint angles analysis
• Model performance metrics
• Evaluation results from different training runs

Dependencies

Data Processing Dependencies

• numpy>=1.19.2
• pandas>=1.2.0
• scipy>=1.6.0
• scikit-learn>=0.24.0
• matplotlib>=3.3.0

Model Dependencies

• All data processing dependencies plus:
• tensorflow>=2.4.0
• joblib>=1.0.0