Pipelines

Module with pipelines for time-series forecasting.

API details

Base:

BasePipeline(horizon)

Base class for pipeline.

Pipelines:

Pipeline(model[, transforms, horizon])	Pipeline of transforms with a final estimator.
AutoRegressivePipeline(model, horizon[, ...])	Pipeline that make regressive models autoregressive.
HierarchicalPipeline(reconciliator, model[, ...])	Pipeline of transforms with a final estimator for hierarchical time series data.

Backtest utilities:

FoldMask(first_train_timestamp, ...)

Container to hold the description of the fold mask.

Pipeline utilities:

assemble_pipelines(models, transforms, horizons)

Create pipelines with broadcasting from models, transforms and horizons.