I have to build a knowledge base of the domain class of insurance (country specific -> Switzerland).
The expected result is the following:
An ontology with classes, subclasses, data properties etc of the class of insurance. It should be possible to feed the ontology with instances of insurance policies (represented as XML-files).
I would like to use Protégé as ontology building tool. I have seen that there is the possibility to import XML-files with the XML Tab function of Protégé which would be useful to import the mentioned data of policies to create instances within the ontology.
My main source of knowledge is a 1'100 page encyclopedia and some policies. As a first step I would like to generate an ontology out of the encyclopedia. As a further step I would refine the ontology manually by using the policies (adding e.g. data properties).
But I am concerned about the encyclopedia and its volume. Manually construct an ontology out of the encyclopedia is not an option due to lack of time.
So my question:
Is there a good practice / tool to build automatically or semi-automatically an ontology out of one large document such as an encyclopedia? I am beginner in text mining or text extraction tasks, but would the programming language "R" be useful? Or rather text-extraction using Apache PDFBox since the encyclopedia is already in a semi-structured form?
Thanks a lot for any suggestions.