Title: Classificação de Textos Através do Modelo Nebuloso de Formas e Métodos de Agregação
Authors: Araújo, Evandro de Oliveira
Abstract: This paper approaches the problem of using pattern recognition techniques in large text databases for documents classification. It is proposed a vector representation for documents based on the frequency of words in the document. This vector representation of the texts corresponds to two-dimensional shapes for which we can determine a fuzzy model. A document can be classified comparing its fuzzy model with the fuzzy models of the prototypes of the available classes. An unknown text belongs to a class whose fuzzy model is the most similar to the fuzzy model of the text. It is also proposed to classify a text based on the idea of clustering. The class of the text is that of the nearest prototype. Some texts have been downloaded from the web and used for classification purposes. The simulation results obtained show the validity of the approach.
Paper as PDF: 6CBRN_039.PDF
BibTex file: 6CBRN_039.bib