BaseRankingEvaluator

class mowl.evaluation.BaseRankingEvaluator(heads, tails, batch_size, device)

Bases: object

Base class for ranking evaluation of ontology embedding methods.

Methods Summary

compute_ranking_metrics(evaluation_model, ...)	Compute the ranking metrics for the evaluation model on the test data.
get_expanded_scores(evaluation_model, batch, ...)
get_scores(evaluation_model, batch)
update_filtering_labels(data)

Methods Documentation

compute_ranking_metrics(evaluation_model, test_data, filter_data=None, mode='head_centric')

Compute the ranking metrics for the evaluation model on the test data. :param evaluation_model: The evaluation model. :type evaluation_model: torch.nn.Module :param test_data: The test data containing the indices of the embeddings :type test_data: torch.Tensor :param filter_data: The filter data containing the indices of the embeddings :type filter_data: torch.Tensor :param mode: The mode of the evaluation. :type mode: str :return: The computed ranking metrics. :rtype: dict

get_expanded_scores(evaluation_model, batch, mode)

get_scores(evaluation_model, batch)

update_filtering_labels(data)