MetricComparison#

class scikitplot.corpus.MetricComparison(metric, candidate, baseline, delta, wins, ties, losses, n_pairs, p_value)[source]#

One metric of an EvaluationComparison.

Parameters:

metricstr

Metric name.

candidate, baselinefloat

Means over the paired queries.

deltafloat

candidate - baseline.

wins, ties, lossesint

Queries where the candidate is higher, equal, lower.

n_pairsint

Queries where both values are defined.

p_valuefloat

Two-sided paired sign-flip randomisation p-value; NaN without pairs.

Parameters:

• metric (str)

• candidate (float)

• baseline (float)

• delta (float)

• wins (int)

• ties (int)

• losses (int)

• n_pairs (int)

• p_value (float)

baseline: float[source]#

candidate: float[source]#

delta: float[source]#

losses: int[source]#

metric: str[source]#

n_pairs: int[source]#

p_value: float[source]#

ties: int[source]#

to_dict()[source]#

Return a JSON-compatible mapping (NaN stays a float).

Return type:

dict[str, Any]

wins: int[source]#