Project

Required Deliverables (for all projects):

• Proposal Presentation (5 March 2008) 5 to 10 minutes per project including discussion
• Paper describing the topic, approach and discussing results and/or suggestions, future work (prefered 6 pages, maximal 8 pages)
• Final Presentation (14 April 2008) 10 to 15 minutes per project including discussion
• Deadline paper: 18 April 2008, 23.59

The aim of the assignment is to get hands-on experience in the development of ontologies and its direct application in the World Wide Web. The description below should therefore be seen more as an overall suggestion than a strict recipe to be followed. As the seminar evolves, we will discuss together how and what to shape in the project. There are some possible topics for the project:

• Course Assistant: Make your own course webpage from the information on the course webpages (UU, ICS). Search by course topic, level, entry requitements, lecturers, schedule. Make sure it is able to alert you for due deadlines, points you to related websites, includes slides and other class information. You will need a course ontology and to use page scrapers such as Piggy_Bank.

• Ontologies and eLearning: use of ontologies in order to structure and retrieve learning material. As and example you can have a look at the project for an electronic book on linguisics for which a small ontology has been built to structure the learning material: www.let.uu.nl/linguistics/log

• Web Shop: In this case study, the catalogues from multiple individual companies should be organized and updated regularly in a unified catalog. This will help users to effectively and transparently access, relate, and combine products from different producers. The problem is that each company has its own catalog organized in a different way. Some companies allow users to browse their catalogues on a product basis (keywords in the product name or description), others provide the user with a nice directory of product categories, families.

• Linguistic corpora: Language resources such as corpora, databases, as well as electronic dictionaries and grammars are being developed and made available to the linguistic community via the World-Wide Web. Quite frequently these resources have been built independently resulting in heterogeneous systems and in severe limits with respect to their appropriate and efficient use. Heterogeneity is found at two different levels: technical level (different hardware platforms, operating systems, etc.) and conceptual level (different data representations and data models for similar objects as well as semantic differences and ambiguities).

  An interesting way to overcome the heterogeneity at the conceptual level makes extensive use of an ontology of linguistic notions. We will develop (fragments) of such an ontology and we will explore whether available language tools can help in the authomatic development of ontologies.

• Language Technology for ontology building: Explore the possibility of using annotated corpora to create ontologies automatically.
  To this end you will need open source tools such as the ones available through the GATE project (cf. link in session 6) as well as corpora. An example of this kind of work is the master thesis by Groenestege (cf. litterature list in session 6). It might also be relevant to have a look at the Onto-LT project (cf. link in session 6).

• Folksonomy vs. Ontologies: Possibility to explore search and retrieval on a given topic by means of tagging and by means of ontologies. For example, consider organizing a collection of music CDs using an ontology (ex. using Protege) or annotation techniques as used in Folksomony.

For all projects, you will problably use Protege. The official webpage of Protege (for free download, tutorials and other information) can be found here.