I thank you for your answer. This is an absolutely interesting question. The method we will use is based on PubMed Entrez API. Using a Python code, we retrieve the number of PubMed publications jointly involving the name of the disease and the name of the risk
factor and we retrieve the PubMed ID of the most relevant search result. As you already know, the relations having important number of query results are more likely to be accurate. After we do this, we will give the list of the retrieved risk factors to a
number of physicians for verification. After that, we will add them to Wikidata using QuickStatements. As for the concerns you raised, they are absolutely accurate. We can discuss them in details after the acceptance of the proposal. You can also be a co-author
of the work.
De : Wikidata <email@example.com> de la part de Egon Willighagen <firstname.lastname@example.org>
Envoyé : lundi 13 août 2018 06:50
À : Discussion list for the Wikidata project.
Cc : email@example.com; firstname.lastname@example.org; email@example.com
Objet : Re: [Wikidata] Please support risk factor Wikidata property
interesting proposal, indeed. I left a comment on the page, which basically says we need qualifiers and references. The references because some claims will be controversial (even if backed up by literature), while I also see these links as likely time
bound, due to changing health policies (I can think of a few other reasons), at least for the residence examples (I hope no one gets any ideas about Q4 'risk factor' 'residence' 'war zone').
With kind regards,