poprox_recommender.evaluation.metrics

Functions

convert_df_to_article_set(rec_df)
measure_user_recs(user)	Measure a single user's recommendations.

Classes

UserRecs(user_id, personalized, recs, truth)

A user's recommendations (possibly from multiple algorithms and stages)

poprox_recommender.evaluation.metrics.rank_biased_overlap(recs_list_a, recs_list_b, p=0.9, k=10)

Computes the RBO metric defined in: Webber, William, Alistair Moffat, and Justin Zobel. “A similarity measure for indefinite rankings.” ACM Transactions on Information Systems (TOIS) 28.4 (2010): 20.

https://dl.acm.org/doi/10.1145/1852102.1852106

Parameters:

• recs_list_a (ArticleSet)

• recs_list_b (ArticleSet)

• p (float)

• k (int)

Return type:

float

class poprox_recommender.evaluation.metrics.UserRecs(user_id, personalized, recs, truth)

Bases: NamedTuple

A user’s recommendations (possibly from multiple algorithms and stages)

Parameters:

• user_id (UUID)

• personalized (bool)

• recs (DataFrame)

• truth (DataFrame)

user_id: UUID

Alias for field number 0

personalized: bool

Alias for field number 1

recs: DataFrame

Alias for field number 2

truth: DataFrame

Alias for field number 3

poprox_recommender.evaluation.metrics.measure_user_recs(user)

Measure a single user’s recommendations. Returns the user ID and a data frame of evaluation metrics.

Parameters:

user (UserRecs)

Return type:

tuple[UUID, DataFrame]

Modules

rbo