ProbedReach#

class pymc_marketing.mmm.incrementality.ProbedReach(lease, baseline, baseline_array, probe, reach, effect_names, trailing_lags, channel_data_dims)[source]#

Leased evaluator, factual baseline and measured reach of one evaluation.

Parameters:
  • lease (EvaluatorLease) – The compiled batched evaluator, held for the duration of the request.

  • baseline (dict) – The factual evaluation of every node, per node (n_samples, n_dates, *non_date_dims).

  • baseline_array (ndarray) – The factual spend the baseline was evaluated on.

  • probe (SpendProbe) – The probe, for what is measured on demand (channel mixing).

  • reach (SpendReach) – How far in time a change in spend moves the evaluated nodes.

  • effect_names (tuple[str, ...]) – The contribution variables of the channel-dependent effects evaluated.

  • trailing_lags (int) – The adstock kernel’s own trailing lags.

  • channel_data_dims (list[str]) – channel_data’s declared axis order.

Methods

ProbedReach.__init__(lease, baseline, ...)

ProbedReach.release()

Return the evaluator to the cache, also when the request raised.

Attributes