Crowdsourcing techniques to create a fuzzy subset of SNOMED CT for semantic tagging of medical documents

Ontologies and other schemes are useful for allowing semantic tagging of documents for many applications on the semantic web. Representing uncertainty on the semantic web is becoming increasingly common, using fuzzy ontologies and other techniques. Very large ontologies and vocabularies have been created, however users may find it difficult to select the correct concept or term when there are large numbers of items that on face value appear to represent the same idea. Creating subsets of ontologies is a popular approach to solving this problem but this may not fit well with the need to deal with complex domains. However crowdsourcing techniques, which harness the power of large groups, may be more effective than document analysis or expert opinion. In Crowdsourcing, large numbers of people collaborate by performing relatively simple tasks usually using applications distributed via the World Wide Web. This approach is being tested in the medical domain using a very large clinical vocabulary, SNOMED CT.