F1Beta

class rectools.metrics.classification.F1Beta(k: int, debias_config: DebiasConfig = None, beta: float = 1.0)[source]

Bases: SimpleClassificationMetric

Fbeta score for k first recommendations. See more: https://en.wikipedia.org/wiki/F-score

The f1_beta equals to (1 + beta_sqr) * p@k * r@k / (beta_sqr * p@k + r@k) where

  • beta_sqr equals to beta ** 2

  • p@k: precision@k equals to tp / k where
    -tp is the number of relevant recommendations

    among first k items in the top of recommendation list.

  • r@k: recall@k equals to tp / liked where
    • tp is the number of relevant recommendations

      among first k items in the top of recommendation list;

    • liked is the number of items the user has interacted

      (bought, liked) with (in period after recommendations were given).

Parameters

  • k (int) – Number of items in top of recommendations list that will be used to calculate metric.

  • beta (float) – Weight of recall. Default value: beta = 1.0

  • debias_config (DebiasConfig, optional, default None) – Config with debias method parameters (iqr_coef, random_state).

Inherited-members

Methods

calc(reco, interactions)

Calculate metric value.

calc_from_confusion_df(confusion_df[, ...])

Calculate metric value from prepared confusion matrix.

calc_per_user(reco, interactions)

Calculate metric values for all users.

calc_per_user_from_confusion_df(confusion_df)

Calculate metric values for all users from prepared confusion matrix.

Attributes

beta