This post is part of my series of post about linked data, in this post I suppose to talk about ontologies in general, but there are a lot of academic papers and thesis about ontologies, in a blog post, I could not go further as them, so I decide to change my approach,  I am going to show a practical example how I see the usage of an ontology based on my own professional experience, so please, allow me to share my view about what a ontology is with you, in my point of view an Ontology is where we define types, properties and relationships of concepts in a thesaurus, I know, this definition to bring more doubts then answer, but hold this idea for a second, I will talk about it later.

The ontology that I am going use as example is the Simple Knowledge Organization System or just (SKOK),  it is defined on its reference page by the W3C as “a common data model for knowledge organization systems such as thesauri, classification schemes, subject heading systems and taxonomies. Using SKOS, a knowledge organization system can be expressed as machine-readable data. It can then be exchanged between computer applications and published in a machine-readable format in the Web.”, thesauri, classification schemes, subject heading systems and taxonomy, you as a librarian must to heard about all of these before, so with SKOS we can define taxonomies in linked data format, machinable readable, and automatically provide disambiguation of concepts, so for example, in the same University the School of Medicine and the Nursing school  can have their own taxonomy, they don’t have to agree about the usage, nor definition, nor labels of a concept anymore, which one of them can have their on taxonomy and the linked part of their data will allow the disambiguation of the concepts.

To clarify what I have just said, please look the image below, each one of the circle represent one concept,  concepts are linked to their properties by a line, of course that the lines are relationships. Actually, these are RDF concepts, hence what we have in here are subject, predicates and objects, if you did not understand what a RDF is, please consider read the “close look at RDF” now. The first thing that we should have in mind is that all circles below are concepts, these concepts can be subjects or objects, the properties of the concepts are objects, the values of properties are not represented in the image, instead you see some URI’s of the SKOS vocabulary, you can find the whole list of SKOS URI’s in the SKOS reference page, the URI’s that I am using here are: skos:prefLabel, skos:altLabel, skos:broader, skos:related, skos:definition, skos:narrow and skos:schema, they are pretty straightforward if you related with the knowledge that you as a librarian have about thesaurus, controlled vocabularies, classification schemes, subject heading and taxonomy.


The picture above is a graph, by definition there is no start point or end point in a graph, but I can use the colors to make my point here, please look the blue circle in the picture, this blue circle is linked to yellows concepts, the yellow circles represent the narrow concepts, hence the blue circle is a broader concept of the yellow concepts. The blue have one broader concept the orange circle, it also has a related concept, the purple circle, the green circle is the schema of the blue circle. In the picture, you can see the “properties” of the concepts, these property are the SKOS URI that I mentioned before such as alternative label (skos:altLabel), preferential label (skos:prefLabel) and definition (skos:definition), however, there are also some SKOS URI that define relation between Concepts: broader (skos:broader) that represents a hierarchical relationship from the concept below to the upper concept, narrower (skos:narrower) that represents a hierarchical relationship from the upper concept to the below concept and related (skos:related) that represent relation between terms related, I meant, no hierarchical, below you can see an turtle file with some concepts:

In the picture above, you can see three concepts, lines 4, 10 and 17, you should remember form the post about RDF that those concepts are represented by URI. In the line 6 you can see an example of alternatives labels, in the line 8 you can see an example of preferential, and you can see also the relations of concepts, note all concept here are RDF concept, so you should see them as Subjects, predicates and objects, so a concept that is a subject in on line can be a predicate in another, that is the case broaders concept in the lines 12 and 19.

Up to here I did a very high level overview about how to use SKOS to describe a taxonomy, but it should be enough for a librarian that have experience with controlled vocabularies make some concetions, it is pretty obvious now the we can use SKOS to generate taxonomies, thesaurus, controlled vocabularies and subjects heading with SKOS, but you should be asking yourself by now: I can do it with the tools that I already have in place, so what is the point here? The simple answer for this question is disambiguation, because the concepts are URI and we can refer then as URI, we can have two institutions referring the same concept with differents preferential and alternative labels, even more, in one taxonomy you could have a concept in a different hierarchical of another one, for example, as I said before, the concept “care” can be have differents labels in a the Nursing School then it has in the Medicine school of the same University or even in others Institutions. To show you guys another usage these thing, I did a quick video with proof of concept using SKOS to create a semantic suggestion to the user in a web interface, it is the first tutorial video that I did myself, I am trying to find out how to do videos, so please be patient.

I hope you guys enjoy this post, and if you any comment or question please, my intention is create a channel among librarians to share knowledge in linked data, so your feedback is very important to me.

Leave a Reply

Your email address will not be published. Required fields are marked *