Article,
Simple Method for Ontology Automatic Extraction from Documents
International Journal of Advanced Computer Science and Applications(IJACSA), (2012)
Abstract
There are many situations where it is needed to represent and analyze the concepts that describe a document or a collection of documents. One of such situations is the information retrieval, which is becoming more complex by the growing number and variety of document types. One way to represent the concepts is through a formal structure using ontologies. Thus, this article presents a fast and simple method for automatic extraction of ontologies from documents or from a collections of documents that is independent of the document type and uses the junction of several theories and techniques, such as latent semantic for the extraction of initial concepts, Wordnet and similarity to obtain the correlation between the concepts.
