Título: Técnicas Baseadas Em Redes Neurais Artificiais E Lógica Fuzzy Para Mineração De Textos
Autores: Mendonça, Jonas Henrique; Drummond, Isabela Neves; Sandri, Sandra Aparecida
Resumo: There is currently a considerable amount of information available in text format on the Internet and the networks of large corporations. This information can be found in an unstructured format, difficult to handle by computer programs. It is proposed in this work, the study of a neuro-fuzzy system applied to the task of text mining. The models investigated for the clustering task of texts is a neural network fuzzy-ART and a neural network SOM. To validate this model experiments are carried out with the bases of texts Reuters Transcribed Subset and SyskillWebert. Finally, the fuzzy-ART network is compared with the neural network SOM.
Palavras-chave: Text mining; fuzzy logic; neural networks; neuro-fuzzy; clustering
Páginas: 8
Código DOI: 10.21528/CBIC2011-27.5
Artigo em pdf: st_27.5.pdf
Arquivo BibTex: st_27.5.bib