search

dice-group / dice-embeddings

Hardware-agnostic Framework for Large-scale Knowledge Graph Embeddings
MIT License
48 stars 13 forks source link

Are the missing triples found by ConEx ranked? #97

Closed hzahera closed 1 year ago

hzahera commented 1 year ago

Hi Demir,

I have a question regardind the output of find_missing_triples function, are the output (i.e. missing triples) ranked according w.r.t the confidence score? Which means the first missing_triple has the highest ranking score? 

from dicee import KGE
pre_trained_kge = KGE(path_of_pretrained_model_dir='ConEx')
missing_triples = pre_trained_kge.find_missing_triples(confidence=0.95, entities=[''], relations=[''])
Demirrr commented 1 year ago

The output is a set of triples, hence, there is no ranking between the returned/found missing triples. Each triple in the returned set is a triple whose predicted likelihood score is greater than the confidence threshold.

Please read the documentation for details.