Friday, 7 August 2015

SKOS and BibTeX in Creating Semantic Ontology on Medical Articles

There are numerous projects that can serve as useful foundations for forming your ontology. One of such projects is Simple Knowledge Organization System (SKOS), a W3C Recommendation.
SKOS provides a standard way to represent knowledge organization systems such as thesauri, classification schemes using the RDF. Another project is BibTeX, a method of marking up bibliographic data, primarily for use in LaTeX documents, but also useful for generic bibliographic storage.

With Fluent Editor, you can utilize both projects through importing them as references, which can be useful for expressing better organization of knowledge. In this post, we will present how SKOS and BibTeX can be utilized in creating your ontology. Data used in the following ontology is based on an excerpt from a list of medical articles on PubMed Central (PMC).

Forming Knowledge Structure
We will walk through a sample ontology 'Medical Article' on Fluent Editor which uses SKOS and BibTeX. Open it on New menu.

In order to use SKOS and BibTeX, first they should be included as references into our ontology file on the bottom.

[bibtex] '' ('')
[skos] '' (''). 

There are a number of journals available on PMC and each journal contains numerous articles. Here in this example, we will express abridged information of journals and articles and some information will be omitted. Let’s start by describing a journal instance that has title and ISSN, which we can adapt from BibTeX standard. For future convenience in reasoning, we can also add two new properties such as write-about and has-concept.
Journal-1 is a journal and has-title[bibtex] equal-to 'ACS Chemical Neuroscience' and has-i-s-s-n[bibtex] equal-to '1948-7193' and has-concept Acs-Chemical-Neuroscience and writes-about equal-to 'ACS Chemical Neuroscience'.
While typing, Hinting Textbox of Fluent Editor can guide you which entities are provides from which references.

Next, some article instances will be added which belong to Journal-1. Each article belongs to a journal and has information such as - PMCID, title, author, concept, volume, and so force. Again, you can see many entities are directly adapted from BibTeX.
Article-1 is an article[bibtex] and has-journal[bibtex] Journal-1 and has-pmcid equal-to 'PMC3894722' and has-title[bibtex] equal-to 'Wavelength-Selective One- and Two-Photon Uncaging of GABA' and has-author[bibtex] equal-to 'Joseph M. Amatrudo' and has-author[bibtex] equal-to 'Jeremy P. Olson' and has-author[bibtex] equal-to 'G. Lur'  and has-author[bibtex] equal-to 'Chiayu Q. Chiu'  and has-author[bibtex] equal-to 'Michael J. Higley'  and has-author[bibtex] equal-to 'Graham C. R. Ellis-Davies' and has-year[bibtex] equal-to 2014 and has-volume[bibtex] equal-to 5 and has-concept Gamma-Aminobutyric-Acid and writes-about equal-to 'ACS Chemical Neuroscience' and writes-about equal-to 'Wavelength-Selective One- and Two-Photon' and writes-about equal-to 'GABA'.
Now pay attention to the property has-concept. While Journal-1 has concept Acs-Chemical-Neuroscience, Article-1 has concept Gamma-Aminobutyric-Acid, which is actually a sub-concept of that of Journal-1’s. In order to express such links between concepts, we can take use of SKOS. Here we use narrowTransitive.
Acs-Chemical-Neuroscience is a concept[skos].
Gamma-Aminobutyric-Acid is a concept[skos] and "narrowerTransitive"[skos] Acs-Chemical-Neuroscience

With this knowledge you can retrieve articles that is writing about subconcept of a given concept, by querying such as : 
"Who-Or-What has-concept that "narrowerTransitive"[skos] Acs-Chemical-Neuroscience?"

You can also make other interesting queries depending on which properties you add more. For example, since we added a property writes-about, we can query : Who-Or-What writes-about equal-to 'GABA'? ...

...or Who-Or-What is an article[bibtex] that has-concept Gamma-Aminobutyric-Acid?  

... and so force.

In conclusion, it is easily achievable to utilize multiple external ontology files using Fluent Editor. Once they are referenced, a HintBox will guide you in real time while writing your knowledge, retrieving all the entities defined in the references. Moreover, reasoning tools will include external entities that are used and reflect them in calculation. Such interfaces will provide you convenience in using standard or popular ontology files such as SKOS and BibTeX, which would consequently enhance knowledge organization and universality of your ontology file.

More Resources:

No comments:

Post a Comment