Model Configuration
Manage the foundational models powering predictive forecasting signals across the restaurant network.
Forecasting Engines
Select the primary model used for daily demand and inventory-depletion forecasting.
Current Active Signal: The global GBM (LightGBM) is trained on ~30 days of historical POS data with local weather overlays to predict daily prep needs.
Configure the secondary model for shift-note parsing, anomaly detection, and waste-reduction insights.