search

KnowledgeCaptureAndDiscovery / wings-site

Repository for WINGS website
https://knowledgecaptureanddiscovery.github.io/wings-site/home.html
1 stars 1 forks source link

Tutorial: types of constraint reasoning #7

Open dgarijo opened 6 years ago

dgarijo commented 6 years ago

I have found out that you can do constraint reasoning without rules. You can add constraint to your workflow such as:

C1: "input1 dcdom:hasLocation ?loc" C2: "input2 dcdom:hasLocation ?loc".

And then the data suggestion feature will only show datasets with the same location :)

We should document this in the tutorial

dgarijo commented 6 years ago

[ backwardClassifierModelAndTestDataHaveSameDomain: print(Firing backwardClassifierModelAndTestDataHaveSameDomain) (?c rdf:type pcdom:ClassifierClass) (?c pc:hasInput ?idvmodel) (?idvmodel pc:hasArgumentID 'model') (?c pc:hasInput ?idvdata) (?idvdata pc:hasArgumentID 'testData') noValue(?idvmodel dcdom:hasDomain) noValue(?idvdata dcdom:hasDomain) uriConcat(?idvmodel '_domain' ?dom) -> (?idvmodel dcdom:hasDomain ?dom) (?idvdata dcdom:hasDomain ?dom) print(Setting domain of ?idvmodel and ?idvdata to ?dom because both inputs to the Classifier should be from the same domain) ]