Sealant

pykeen.triples.leakage Module

Tools for removing the leakage from datasets.

Leakage is when the inverse of a given training triple appears in either the testing or validation set. This scenario generally leads to inflated and misleading evaluation because predicting an inverse triple is usually very easy and not a sign of the generalizability of a model to predict novel triples.

Functions

unleak(train, *triples_factories[, n, ...])

Unleak a train, test, and validate triples factory.

reindex(*triples_factories)

Reindex a set of triples factories.

Classes

Sealant(triples_factory[, ...])

Stores inverse frequencies and inverse mappings in a given triples factory.

Class Inheritance Diagram

Inheritance diagram of pykeen.triples.leakage.Sealant