Populating the Ontology with Instances

In order to populate the ontology with individuals, a bibliography database was initially obtained from VU's Pure service \cite{pure}. Due to database being in .bib format, and due to lack of sufficiently high quality .bib to RDF conversion packages in Python, a parser and .ttl converter was programmed using Python (and 'pybtex' package ). Although the data preparation part of the ontology creation process had cost most of the available time in this way, the time investment was seen as necessary due to Pure dataset being an important element for the project. These scripts can be reached at 'https://github.com/clokman/KAD/tree/master/Prototype_Final'. The reader is encouraged to view the Python scripts as they would likely be able to demonstrate a good degree of understanding of the concepts of the course and may be relevant for evaluation (related files will also be attached together with the submission of this report).
It should also be noted that due to practical concerns (i.e., lack of time and computing power), a truncated version of the actual Pure database was used for importing instances, and the number of instances imported to the ontology was limited to roughly 375 instances (124 KB in size). The original Pure bibliography contains more than 1.5 million lines (about 100 MB), and it will be used later in the project when more time and computational resources are available.
Each instance in the .bib file consisted of bibliographic information regarding to a document, and had varying elements (e.g., while some publication instances had a lot of detail, some only had the basic information such as author, title and year). Therefore, the parser script was programmed to be able to deal with missing variables on a regular basis.

Adding External Classes

Although a few example scientific domains (e.g., computer science) were used as placeholders during previous assignments, a comprehensive and accurate domain map of scientific fields were necessary for the purposes of the overarching project (although less necessary for the current stage of the project and the course). Therefore,  as a convenient and trustworthy way of categorizing fields of science, ' Web Of Science Category Terms' \cite{reuters2017} was added to the ontology as classes through parsing and conversion to .ttl.  (This Python script is named 'd_web-of-science-categories.py' in the GitHub repository and is also among submission files.)

Conceptualization and Realization

The current version of the ontology includes 28 classes for describing the domain of scientific publications and academic research output in general:
Through using properties (fig. \ref{797019}), a conceptual network between these classes has been established. 39 object properties describe:
A visual summary and explanations of the structure of the ontology can be seen in figures \ref{675710} through \ref{797019}.