UnrepeatedReco

class rectools.metrics.dq.UnrepeatedReco(k: int, deep: bool = False)

Bases: _RecoDQMetric

Unrepeated items recommended to the same user in recommendations table. This metrics help to identify situations when recommendation lists have duplicated items for same users. Specify deep=False to calculate share of user without any duplicated itemd at first k positions. Specify deep=True to calculate average share of unrepeated items for each user at first k positions.

Parameters

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

• deep (bool, default False) – Whether to calculated detailed value of the metric for each user. Otherwise just the share of users without identified problem will be returned (this is the default behaviour).

Examples

>>> reco = pd.DataFrame(
...     {
...         Columns.User: [1, 1, 2, 2, 2, 3, 3, 3, 3, 3],
...         Columns.Item: [1, 2, 1, 1, 3, 1, 2, 2, 1, 5],
...         Columns.Rank: [1, 2, 1, 2, 3, 1, 2, 3, 4, 5],
...     }
... )
>>> UnrepeatedReco(k=1).calc_per_user(reco).values
array([1, 1, 1])
>>> UnrepeatedReco(k=4).calc_per_user(reco).values
array([1, 0, 0])
>>> UnrepeatedReco(k=4, deep=True).calc_per_user(reco).values
array([1.        , 0.66666667, 0.5       ])

Inherited-members

Parameters

• k (int) –

• deep (bool) –

Methods

calc(reco)	Calculate metric value.
calc_per_user(reco)	Calculate metric values for all users.

Attributes

calc_per_user(reco: DataFrame) → Series

Calculate metric values for all users.

Parameters

reco (pd.DataFrame) – Recommendations table with columns Columns.User, Columns.Item, Columns.Rank.

Returns

Values of metric (index - user id, values - metric value for every user).

Return type

pd.Series