. . . "Learning of N-layers neural network"@en . "U\u010Den\u00ED n-vrstv\u00E9 neuronov\u00E9 s\u00EDt\u011B"@cs . "43110" . "back-propagation error; handling of learning rate; learning rate; learninng algoritm; matrix form; neural networks"@en . "U\u010Den\u00ED n-vrstv\u00E9 neuronov\u00E9 s\u00EDt\u011B" . "Mati\u00E1\u0161ov\u00E1, And\u011Bla" . "R\u00E1bov\u00E1, Ivana" . "6" . . . "LIII" . . . . "V posledn\u00EDch desetilet\u00EDch lze zaznamenat zna\u010Dn\u00FD n\u00E1r\u016Fst aplikac\u00ED pro \u0159e\u0161en\u00ED \u00FAloh z r\u016Fzn\u00FDch oblast\u00ED lidsk\u00E9 \u010Dinnosti vyu\u017E\u00EDvaj\u00EDc\u00ED technologie um\u011Bl\u00E9 inteligence. Z\u00E1jem o tyto technologie lze p\u0159ipsat skute\u010Dnosti, \u017Ee klasick\u00E9 zp\u016Fsoby \u0159e\u0161en\u00ED bu\u010F neexistuj\u00ED nebo pro svoji robustnost nejsou vhodn\u00E9. \u010Casto jsou vyu\u017E\u00EDv\u00E1ny v aplikac\u00EDch %22Business Inteligence%22 umo\u017E\u0148uj\u00EDc\u00EDch z\u00EDsk\u00E1vat pot\u0159ebn\u00E9 informace pro kvalitn\u00ED rozhodov\u00E1n\u00ED a zvy\u0161ov\u00E1n\u00ED konkurenceschopnosti. Jedn\u00EDm z nejroz\u0161\u00ED\u0159en\u011Bj\u0161\u00EDch n\u00E1stroj\u016F um\u011Bl\u00E9 inteligence jsou v\u00EDcevrstv\u00E9 neuronov\u00E9 s\u00EDt\u011B. Jejich velkou v\u00FDhodou je relativn\u00ED jednoduchost a mo\u017Enost samou\u010Den\u00ED na z\u00E1klad\u011B souboru vzorov\u00FDch situac\u00ED. Pro etapu u\u010Den\u00ED se nej\u010Dast\u011Bji pou\u017E\u00EDv\u00E1 algoritmus zp\u011Btn\u00E9ho \u0161\u00ED\u0159en\u00ED chyby. Av\u0161ak p\u0159i jeho realizaci a po\u010D\u00E1te\u010Dn\u00EDm vyu\u017E\u00EDv\u00E1n\u00ED zjist\u00EDme, \u017Ee je nutn\u00E9 jej doplnit vhodn\u00FDm zp\u016Fsobem \u0159\u00EDzen\u00ED velikosti koeficientu u\u010Den\u00ED, na jeho\u017E volb\u011B je z\u00E1visl\u00E1 konvergence procesu u\u010Den\u00ED. C\u00EDlem toho p\u0159\u00EDsp\u011Bvku je kompaktn\u00ED vyj\u00E1d\u0159en\u00ED algoritmu u\u010Den\u00ED v maticov\u00E9m tvaru, odvozen\u00ED metody \u0159\u00EDzen\u00ED koeficie"@cs . "Kone\u010Dn\u00FD, Vladim\u00EDr" . "3"^^ . . . . "In the last decade we can observe increasing number of applications based on the Artificial Intelligence that are designed to solve problems from different areas of human activity. The reason why there is so much interest in these technologies is that the classical way of solutions does not exist or these technologies are not suitable because of their robustness. They are often used in applications like Business Intelligence that enable to obtain useful information for high quality decision making and to increase competitive advantage. One of the most widespread tools for the Artificial Intelligence are the artificial neural networks. Their high advantage is relative simplicity and the possibility of self-learning based on set of pattern situations. For the learning phase is the most commonly used algorithm back-propagation error (BPE). The base of BPE is the method minima of error function representing the sum of squared errors on outputs of neural net, for all patterns of the learning set. Howe"@en . "Acta Universitatis Agriculturae et Silviculturae Mendelianae Brunensis" . "3"^^ . . "CZ - \u010Cesk\u00E1 republika" . "RIV/62156489:43110/05:00007456" . . "75;84" . . . "Learning of N-layers neural network"@en . . "547613" . "10"^^ . . "RIV/62156489:43110/05:00007456!RIV06-MSM-43110___" . . "U\u010Den\u00ED n-vrstv\u00E9 neuronov\u00E9 s\u00EDt\u011B" . "Z(MSM6215648904)" . . . "V posledn\u00EDch desetilet\u00EDch lze zaznamenat zna\u010Dn\u00FD n\u00E1r\u016Fst aplikac\u00ED pro \u0159e\u0161en\u00ED \u00FAloh z r\u016Fzn\u00FDch oblast\u00ED lidsk\u00E9 \u010Dinnosti vyu\u017E\u00EDvaj\u00EDc\u00ED technologie um\u011Bl\u00E9 inteligence. Z\u00E1jem o tyto technologie lze p\u0159ipsat skute\u010Dnosti, \u017Ee klasick\u00E9 zp\u016Fsoby \u0159e\u0161en\u00ED bu\u010F neexistuj\u00ED nebo pro svoji robustnost nejsou vhodn\u00E9. \u010Casto jsou vyu\u017E\u00EDv\u00E1ny v aplikac\u00EDch %22Business Inteligence%22 umo\u017E\u0148uj\u00EDc\u00EDch z\u00EDsk\u00E1vat pot\u0159ebn\u00E9 informace pro kvalitn\u00ED rozhodov\u00E1n\u00ED a zvy\u0161ov\u00E1n\u00ED konkurenceschopnosti. Jedn\u00EDm z nejroz\u0161\u00ED\u0159en\u011Bj\u0161\u00EDch n\u00E1stroj\u016F um\u011Bl\u00E9 inteligence jsou v\u00EDcevrstv\u00E9 neuronov\u00E9 s\u00EDt\u011B. Jejich velkou v\u00FDhodou je relativn\u00ED jednoduchost a mo\u017Enost samou\u010Den\u00ED na z\u00E1klad\u011B souboru vzorov\u00FDch situac\u00ED. Pro etapu u\u010Den\u00ED se nej\u010Dast\u011Bji pou\u017E\u00EDv\u00E1 algoritmus zp\u011Btn\u00E9ho \u0161\u00ED\u0159en\u00ED chyby. Av\u0161ak p\u0159i jeho realizaci a po\u010D\u00E1te\u010Dn\u00EDm vyu\u017E\u00EDv\u00E1n\u00ED zjist\u00EDme, \u017Ee je nutn\u00E9 jej doplnit vhodn\u00FDm zp\u016Fsobem \u0159\u00EDzen\u00ED velikosti koeficientu u\u010Den\u00ED, na jeho\u017E volb\u011B je z\u00E1visl\u00E1 konvergence procesu u\u010Den\u00ED. C\u00EDlem toho p\u0159\u00EDsp\u011Bvku je kompaktn\u00ED vyj\u00E1d\u0159en\u00ED algoritmu u\u010Den\u00ED v maticov\u00E9m tvaru, odvozen\u00ED metody \u0159\u00EDzen\u00ED koeficie" . "[3C108AFB1BFF]" . "1211-8516" . "U\u010Den\u00ED n-vrstv\u00E9 neuronov\u00E9 s\u00EDt\u011B"@cs .