langfair.metrics.recommendation.metrics.prag.PRAG#

class langfair.metrics.recommendation.metrics.prag.PRAG#

Bases: Metric

__init__()#

This class computes Pairwise Ranking Accuracy Gap (PRAG).

Methods

__init__()

This class computes Pairwise Ranking Accuracy Gap (PRAG).

evaluate(list1, list2)

This method computes PRAG score between two lists.
evaluate(list1, list2)#

This method computes PRAG score between two lists.

Parameters:

  • list1 (list of strings) – A list of recommendation from an LLM model.

  • list2 (list of strings) – Another list of recommendation from an LLM model.

Returns:

PRAG metric value for the two provided lists of recommendations (float)

Return type:

float