Beyond tree cover: mapping the world’s forest types
deepmind.google·4d ago
AI Engineering
Hugging Face Blog·5d ago·5 min read
Ensemble forecasting with IBM Granite time-series models
TL;DR
IBM Granite introduces an ensemble forecasting approach that combines four pre-trained time-series models, simplifying the selection process for users. This method provides median and quantile forecasts without the need for retraining.
✦ Why It Matters
Start using the Granite TSFM ensemble today to streamline your time-series forecasting process.
Key Takeaways
- Install Granite TSFM to access the ensemble models.
- Use the Probabilistic Ensemble for a straightforward forecasting approach.
- Experiment with the IQR Ensemble for adaptive weighting based on forecast spread.
- Backtest the ensemble on your data to evaluate performance.
How It Works
The ensemble combines four different architectures—Patched Transformer, Patched Conformer, State-space model, and Tiny Time Mixers—each contributing unique insights into the forecast. By pooling their outputs, the ensemble mitigates the risk of relying on a single model, enhancing overall prediction accuracy.