Application of Neural Network Boolean Factor Analysis Procedure to Automatic Conference Papers Categorization

0390803 - ÚI 2013 RIV PT eng C - Conference Paper (international conference)
Húsek, Dušan - Frolov, A. A. - Polyakov, P.Y. - Řezanková, H. - Snášel, V.
Application of Neural Network Boolean Factor Analysis Procedure to Automatic Conference Papers Categorization.
Bulletin of the International Statistical Institute. Lisabon: Instituto Nacional de Estatística, 2008 - (Gomes, M.; Pinto Martins, J.; Silva, J.), s. 3739-3742. ISBN 978-972-673-992-0.
[ISI 2007. Session of the International Statistical Institute /56./. Lisboa (PT), 22.08.2007-29.08.2007]
R&D Projects: GA AV ČR 1ET100300414
Grant - others:RFBR(RU) 05-07-90049
Institutional research plan: CEZ:AV0Z10300504
Keywords : Boolean factor analysis * document classification * automatic concepts search * unsupervised learning * neural network
Subject RIV: BB - Applied Statistics, Operational Research

The neural network algorithm is proposed for automatic unsupervised words categorization using purely statistic information derived from textual data. The method is an extension of the of Boolean factor analysis algorithm (A. Frolov at al., Boolean factor analysis by attractor neural network", IEEE Trans. on Neural Networks, 18, (3), 2007). We apply the method to two types of textual data on Neural Networks. The first data set consists of the papers published in the proceedings of the IJCNN 2003 and 2004 conferences, the second - consists of the papers published in the proceedings of Russian conference on “NEUROINFORMATICS 2004 and 2005.
Permanent Link: http://hdl.handle.net/11104/0219634