Master Thesis Proposal: Ontology Evolution for Dynamic Prediction of Infectious Diseases
During the past decade, the Semantic Web has emerged as an evolution of the human readable web, towards a web of macine processable data with formal semantics. To interpret the semantics of online resources, ontologies are used to model the domain of interest. One application of such ontologies is to interpret and classify the content and type of documents. However, in order to do that in an accurate manner, ontologies need to be up-to-date and represent the current situation in the world, rather than the situation when the ontology was initially constructed. To perform the updating of an ontology in a (semi-)automatic fashion is what is called "ontology evolution".
Web usage statistics have been proven useful in order to predict outbreaks of infectious diseases (c.f. Google flu trends and similar applications). Results are particularly promising when looking at web statistics of official information from healthcare providers, such as the Swedish website 1177.se. However, setting up a static system that analyses the statistics of pages with fixed URLs make the system unable to adapt to changes in the website structure, e.g. changing the URL scheme, and unable to react to changes in user behaviour or the evolution of new keywords. In order to cater for such changes, the system needs to be more flexible, relying on an evolving model of the world (an evolving ontology), rather than hard-coded URL:s and keyword lists.
How to express interest in this topic:
In order to make sure that the student will be able to complete thesis, I usually ask the student to do a small "test" before we agree on starting up the thesis. This is not an exam, just a test to see how quickly the student is able to get into the field. It is also an opportunity for the student to learn more about the technologies involved and see if you are really interested in them, before you start!
So if you are interested in this topic, please access this page and follow the instructions (i.e. fill out the questionnaire and do the small modelling task).
Note that you are free to find any online reading material you like, and use any tool of your choice. But please do not spend more than a day or so (spread out over one week) on this task, so as not to waste too much time if you in the end find it too difficult or not interesting, and in order to let me see how fast you catch on to the techniques and languages.
Don't forget to send me an e-mail expressing your interest in the thesis topic, including the solution to the modelling task, when you are done!
Page responsible: Eva Blomqvist
Last updated: 2013-01-17