tfc-t0
A probabilistic time-series forecasting model that uses historical data and known future variables.
SITE CAPTUREBased on information published on the service's website. Selection and review policy
WHAT IT DOES
What does it do?
tfc-t0 produces probabilistic time-series forecasts using past and known future covariates. predict accepts NumPy arrays, and missing observations can be marked with NaN. You can also construct TimeSeries objects to group inputs of different lengths or use forward to run computations for fine-tuning.
KEY FEATURES
Key features
- Generate probabilistic forecasts using past values and known future covariates
- Accept NumPy arrays as input to predict
- Treat NaN as a missing observation
- Group inputs of different lengths with TimeSeries
- Provide forward as an entry point for fine-tuning