| Resumen |
Knowledge scattered through the Web inside unstructured documents (text documents) can not be easily interpreted by computers. To do so, knowledge contained from them must be extracted by a parser or a person and poured into a suitable data structure, the best form to do this, are with ontologies. For an appropriate merging of these "individual" ontologies, we consider repetitions, redundancies, synonyms, meronyms, different level of details, different viewpoints of the concepts involved, and contradictions, a large and useful ontology could be constructed. This paper presents OM algorithm, an automatic ontology merger that achieves the fusion of two ontologies without human intervention. Through repeated application of OM, we can get a growing ontology of a knowledge topic given. Using OM we hope to achieve automatic knowledge acquisition. There are two missing tasks: the conversion of a given text to its corresponding ontology (by a combination of syntactic and semantic analysis) is not yet automatically done; and the exploitation of the large resulting ontology is still under development.
|
