AuroraAI is a Machine Learning ecosystem designed to bridge heliophysics and predictive analytics. The project is divided into two distinct phases, moving from binary event classification to temporal regression.
The repository is organized into two primary research modules:
- Focus: Determining the probability of an Aurora event (Yes/No) based on real-time solar wind vectors.
- Goal: Establish a high-precision baseline for geomagnetic storm detection.
- Tech: Scikit-Learn, Random Forests, XGBoost.
- Focus: Predicting the specific magnitude of the Kp-Index over a 6-hour horizon.
- Goal: Solve the "intensity" problem using deep temporal architectures.
- Tech: PyTorch, LSTM (Long Short-Term Memory), Temporal Fusion Transformers.
- Python 3.9+
- Pandas / NumPy
- Matplotlib / Seaborn
- NASA/NOAA API Access
Kindly refer to the Notes folder to understand more on this topic and what was used for each module.