Overview of a semantic disambiguation method for unstructured web contexts

摘要

In this paper we give an overview of a multiontology disambiguation method, targeted to discover the intended meaning of words in unstructured web contexts. It receives an ambiguous keyword and its context words as input and provides a list of possible senses for the keyword, scored according to the probability of being the intended one. It accesses any pool of online ontologies as source of word senses, in addition to other available resources. This method is targeted to be used in unstructured contexts that lack well-formed sentences, such as user keywords or folksonomy tags.