0452538 20240103211432.620160215235959.9 84949894891 000391072100022 10.1007/978-3-319-26393-9_22 18 s. P cav_un_epca*0453648978-3-319-26393-9365-382ChamSpringer2016 Product mixtures Mixtures of Dependence Trees EM algorithm cav_un_auth*0101091 Department of Pattern Recognition 80 Grim Jiří UTIA-B Rozpoznávání obrazu Department of Pattern Recognition RO RO Ústav teorie informace a automatizace AV ČR, v. v. i. cav_un_auth*0021092 20 Pudil P. CZ http://library.utia.cas.cz/separaty/2016/RO/grim-0452538.pdf cav_un_auth*0308953 GAP403/12/1557 GA ČR CZ cav_un_auth*0303412 GA14-02652S GA ČR CZ Mixtures of product components assume independence of variables given the index of the component. They can be efficiently estimated from data by means of EM algorithm and have some other useful properties. On the other hand, by considering mixtures of dependence trees, we can explicitly describe the statistical relationship between pairs of variables at the level of individual components and therefore approximation power of the resulting mixture may essentially increase. However, we have found in application to classification of numerals that both models perform comparably and the contribution of dependence-tree structures to the log-likelihood criterion decreases in the course of EM iterations. Thus the optimal estimate of dependence-tree mixture tends to reduce to a simple product mixture model. cav_un_auth*0324200 IJCCI 2014 - International Joint Conference on Computational Intelligence (Rome/Italy) 22.10.2014-24.10.2014 Rome IT BD 2016 2 PR RVO:67985556 http://hdl.handle.net/11104/0257056 cav_un_auth*0324201 Faculty of Management, Prague University of Economics FMJH VSE, Jindřichův Hradec CZ S 3+4 Proceedings Paper Computer Science Artificial Intelligence 2016 84949894891 SCOPUS 000391072100022 WOS cav_un_epca*0453648 Computational Intelligence 978-3-319-26393-9 365 382 Cham Springer 2016 Studies in Computational Intelligence 620