get_corpus_scorer#

scikitplot.corpus.get_corpus_scorer(name, *, response_method='predict', id_getter=None, on_degraded='raise', **kwargs)[source]#

Return a new scorer for a metric name.

Parameters:

namestr

A name from get_corpus_scorer_names; ranking names take an optional cutoff, "recall@10". A ranking name without a cutoff uses the whole ranking.

response_method{“predict”, “search”}, default=”predict”

See make_corpus_scorer.

id_gettercallable or None, default=None

See make_corpus_scorer.

on_degraded{“raise”, “ignore”}, default=”raise”

See make_corpus_scorer.

**kwargsdict

Further arguments of the metric, such as zero_division.

Returns:

scorer

A new instance on every call, as sklearn.metrics.get_scorer.

Raises:

ValueError

For an unknown name, or k given both in the name and as a keyword.

Parameters:

• name (str)

• response_method (str)

• id_getter (Callable[[Any], Any] | None)

• on_degraded (str)

• kwargs (Any)

Return type:

Any

Examples

Try it in your browser!

>>> get_corpus_scorer("ndcg@10", response_method="search")
make_corpus_scorer(ndcg_at_k, response_method='search', k=10)

Go BackOpen In Tab