config_energy.yaml

# seed for the random number generators, to make things reproducible
seed: 0
true_energy_column: corsika_event_header_total_energy

energy:
  #the classifier to use
  regressor : |
    ensemble.RandomForestRegressor(
      n_estimators=30,
      max_features='sqrt',
      n_jobs=-1,
      max_depth=12,
    )

  # randomly sample the data if you dont want to use the whole set
  n_signal: 500

  # define th number of cross validations to perform
  n_cross_validations: 5

  # define the name of the regression target
  target_column: corsika_event_header_total_energy


  # Define the name of the variable you want estimate by regression.
  # The default is 'gamma_energy_energy'.
  output_name: gamma_energy_prediction

  features:
    - size
    - length
    - width
    - num_islands
    - leakage1
    - leakage2
    - skewness_long
    - skewness_trans
    - concentration_cog
    - concentration_core

  # Generate some features using pd.DataFrame.eval
  # List all keys that have to be read from the input files
  # in needed keys.
  # features must be a mapping of feature name -> expression
  feature_generation:
    needed_columns:
      - width
      - length
    features:
      area: width * length * @pi