four_theta.yaml

# @package _global_

defaults:
  - base_local

model:
  _target_: darts.models.forecasting.theta.FourTheta
  #season_mode:  # Note that the default season_mode = MULTIPLICATIVE requires data > 0, thus one must change the data pipeline as below or use another season_mode
  #  _target_: darts.utils.utils.SeasonalityMode
  #  value: "additive"
  normalization: False

datamodule:
  processing_pipeline:
    _target_: darts.dataprocessing.Pipeline
    transformers:
      - _target_: darts.dataprocessing.transformers.boxcox.BoxCox
        lmbda: 0
      - _target_: darts.dataprocessing.transformers.missing_values_filler.MissingValuesFiller
        fill: "auto" # The default, will use pandas.Dataframe.interpolate()