One problem with search is the fact that language is so ambiguous. This combined with the overabundance of information helps library types, and the TKI meta data project, justify their existance by saying that human intervention is needed to remove the ambiguity by hand classifying individual internet resources, and thereby aiding retrieval.
What a small team at Waikato have been doing though is using Wikipedia to do the disambiguation and make the connections. As they put it:
“Wikipedia represents a giant multilingual database of concepts and semantic relations”
One part of their project, Wikipedia Miner does the following:
- provides simplified, object-oriented access to Wikipedia’s structure and content.
- measures how terms and concepts in Wikipedia are connected to each other.
- detects and disambiguates Wikipedia topics when they are mentioned in other documents.
There other project Hopara uses this information to create a browse enhanced search interface into Wikipedia.
Not the answer to all the world’s problems, but huge potential here to start making sense of the web without having to pour millions more into dublin core based meta data projects …. maybe.
