0350163 20240103194203.020101129235959.9 000290245400004 10.1007/978-3-642-15825-4_4 10 s. cav_un_epca*0350162978-3-642-15818-6Part III31-40Berlin HeidelbergSpringer Verlag2010DiamantarasK.DuchWlodzislawIliadisL.S. Probabilistic neural networks Statistical pattern recognition Subspace approach Overfitting reduction cav_un_auth*0101091 Grim Jiří Rozpoznávání obrazu Department of Pattern Recognition RO RO UTIA-B Department of Pattern Recognition Ústav teorie informace a automatizace AV ČR, v. v. i. cav_un_auth*0230019 Hora Jan Rozpoznávání obrazu Department of Pattern Recognition RO RO UTIA-B Ústav teorie informace a automatizace AV ČR, v. v. i. http://library.utia.cas.cz/separaty/2010/RO/grim-computational properties of probabilistic neural networks.pdf GA102/07/1594 GA ČR cav_un_auth*0228611 2C06019 GA MŠk CZ cav_un_auth*0216518 1M0572 GA MŠk cav_un_auth*0001814 CEZ:AV0Z10750506 We discuss the problem of overfitting of probabilistic neural networks in the framework of statistical pattern recognition. The probabilistic approach to neural networks provides a statistically justified subspace method of classification. The underlying structural mixture model includes binary structural parameters and can be optimized by EM algorithm in full generality. Formally, the structural model reduces the number of parameters included and therefore the structural mixtures become less complex and less prone to overfitting. We illustrate how recognition accuracy and the effect of overfitting is influenced by mixture complexity and by the size of training data set. cav_un_auth*0262947 ICANN 2010. International Conference on Artificial Neural Networks /20./ Thessaloniki 15.09.2010-18.09.2010 GR 2011 IN http://hdl.handle.net/11104/0190237 2010 000290245400004 WOS cav_un_epca*0350162 Artificial Neural Networks – ICANN 2010 Part III 978-3-642-15818-6 31 40 Berlin Heidelberg Springer Verlag 2010 Lecture Notes in Computer Science Volume 6354 LNCS Diamantaras K. 340 Duch Wlodzislaw 340 Iliadis L.S. 340