Re: [Wikidata] Enriching Wikidata with Portuguese lexemes

I have matched some Princeton WordNet (PWN) synsets with Wikidata items and some DanNet (Danish wordnet) words with Wikidata lexemes. For PWN, I have used the P2888 with the PWN 3.0 URI (because of the relevance for ImageNet). For PWN synsets-Wikidata items, I have experienced that it is not always trivial to match items [1]. It means that one should be careful with mass upload. For DanNet, I have used the dedicated property P6140. Currently, I have only used DanNet's "words" identifiers which are almost alway trivial to match with Wikidata lexemes and so far not done any mass upload from DanNet data. To me it seems that linking to DanNet 'word' identifiers is ok as the property name itself makes it obvious that the identifier comes from DanNet. A question is if the more semantic information, e.g., hypernyms and synonyms information and perhaps even wordsense information, can be used. I have noted that a lot of Wikidata items have been matched to BabelNet with P2581. BabelNet has a fairly restrictive Creative Commons license. Nevertheless there are close to 60,000 BabelNet identifiers in Wikidata. I wonder if there has been any outreach to BabelNet. [1] http://www2.compute.dtu.dk/pubdb/views/edoc_download.php/7090/pdf/imm7090.p… On 4/13/19 15:06, Houcemeddine A. Turki wrote: