tulip-package: tulip: Robust Probabilistic Forecasts to Tinker With

tulip-packageR Documentation

tulip: Robust Probabilistic Forecasts to Tinker With

Description

logo

An implementation of robust exponential smoothing models that produce probabilistic forecasts using sample paths. Handling of anomalies is integrated into a custom maximum-a-posteriori fitting procedure that exposes most aspects of the model to the user. Initial states can be provided to start training from previous fits, transfer seasonality from other time series. User-defined parameter grids and prior distributions restrict the set of reasonable model choices to robust ones.

Author(s)

Maintainer: Tim Radtke tim@timradtke.net

See Also

Useful links:


timradtke/heuristika documentation built on April 24, 2023, 1:55 a.m.