The DW-NB dataset

[Born1997]

Hello, the following error occurs when I am trying to reproduce AttrE model with DW-NB dataset :

http://dbpedia.org/resource/Marie_d'Alen\u00E7on does not look like a valid URI, trying to serialize this will break.
Traceback (most recent call last):
File "AttrE1.py", line 403, in
if (entity_vocab[URIRef(s.replace('<', '').replace('>', ''))] in entity_kb1_vocab) and (
KeyError: rdflib.term.URIRef("http://dbpedia.org/resource/Marie_d'Alen\\u00E7on")

What does a string like “\u00E7” mean?
Thanks.

[ruizhang-ai]

Hi,
It is a Unicode for character 'ç'. To process the dataset you need to handle the Unicode character.

[Born1997]

Thank you for your reply!
I tried to reproduce the attre model using the DWY-NB dataset, but the results were not ideal. Can you please share the processed pickle file？
Thanks！

[oaimli]

Hi, we have added the DW-NB preprocessed data at: https://bitbucket.org/bayudt/kba/src/master/data/data-wd.zip