Retrieve common entities between wn18rr and FB15k237 datasets

[raphaelsty]

I need to retrieve common entities between both files:

• Entities fb15k237

• Entities wn18rr

The best solution in my opinion is to add Part-of-speech tagging to the entities of the fb15k237 dataset.

I need to retrieve common relations between both files:

• Relations fb15k237

• Relations wn18rr

The relations of the datasets fb15k237 and wn18rr are quite distinct symbolically nevertheless there is a semantic link between the relations of the two datasets.

[raphaelsty]

There are 1552 common entities between WN18RR and FB15K237.

WN18RR shares 3.79% of its entities with FB15K237.

FB15K327 shares 10.88% of its entities with WN18RR.

[raphaelsty]

In FB15K237, there are 13411 training samples (e_i, r, e_j) where e_i and e_j are defined in WN18RR. It represent 4.92% of the training set of FB15K237.

• 4.69% for the valid set of FB15K237

• 4.59% for the test set of FB15K237

[raphaelsty]

I uploaded a new version of wn18rr and fb15k237.

I have aligned the entities of fb15k237 with the entities of wn18rr and I have concatenated the set of relations of both datasets.

New fb15k237

New wn18rr