In Linked Open Data there are many links between the resources and some resources are linked directly and others are linked indirectly through a third resource. some similarity measures of resources depend on a number of direct and indirect links and I am asking If there is any evidence or justification to depend on to give more importance (weight) to direct links over indirect links to find the similarity.
Could you give an example of direct and indirect links?
There are already initiatives to express some kind of similarity, an example could be mapping properties with skos:closeMatch, skos:exactMatch, owl:sameAs, owl:equivalentClass, owl:equivalentProperty (https://www.w3.org/TR/skos-reference/#L4858)
I suggest always keeping in mind the Leibniz law, which I simplify in the following sentence "An object/resource is similar to another object as much they share properties" (Article The Identity of Indiscernibles
).Therefore, for example, given two resources mapping 90% of properties between them, you can say they are likely 90% similar.
I hope it helps you.
Andre Valdestilhas Thank you for the answer. the properties that you mentioned are not always available.
An example to my question, If I need to know the related or similar movies to the initial movie
there are direct links by WikiPageWikiLink property to other movies and other indirect links via a middle resource. so maybe give importance to existing of such direct links over indirect links better than depending on both links equally.
thanks.
regards.
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Graphs