0519831 20241106135817.620200115235959.9 85075330445 000497538000001 10.1080/03081079.2019.1692004 24 s. P cav_un_epca*02567940308-1079491 (2020)88-111Taylor & Francis Bayesian networks Computerized adaptive testing Parameter learning cav_un_auth*0329423 Plajner Martin UTIA-B Matematická teorie rozhodování Department of Decision Making Theory MTR MTR Department of Decision Making Theory CZ Ústav teorie informace a automatizace AV ČR, v. v. i. cav_un_auth*0101228 Vomlel Jiří UTIA-B Matematická teorie rozhodování Department of Decision Making Theory MTR MTR Department of Decision Making Theory Ústav teorie informace a automatizace AV ČR, v. v. i. PDF 2.8 MB http://library.utia.cas.cz/separaty/2020/MTR/plajner-0519831.pdf https://www.tandfonline.com/doi/full/10.1080/03081079.2019.1692004 GA16-12010S GA ČR CZ cav_un_auth*0332303 GA19-04579S GA ČR CZ cav_un_auth*0380558 SGS17/198/OHK4/3T/14 ČVUT CZ cav_un_auth*0361640 Learning parameters of a probabilistic model is a necessary step in machine learning tasks. We present a method to improve learning from small datasets by using monotonicity conditions. Monotonicity simplifies the learning and it is often required by users. We present an algorithm for Bayesian Networks parameter learning. The algorithm and monotonicity conditions are described, and it is shown that with the monotonicity conditions we can better fit underlying data. Our algorithm is tested on artificial and empiric datasets. We use different methods satisfying monotonicity conditions: the proposed gradient descent, isotonic regression EM, and non-linear optimization. We also provide results of unrestricted EM and gradient descent methods. Learned models are compared with respect to their ability to fit data in terms of log-likelihood and their fit of parameters of the generating model. Our proposed method outperforms other methods for small sets, and provides better or comparable results for larger sets. WOS IN 10000 10200 10201 2021 2 4 A sml 4as 20241106135817.6 RVO:67985556 http://hdl.handle.net/11104/0304816 S Publishing Agreement 20191111 3+4 Article Computer Science Theory Methods|Ergonomics C COMPUTERSCIENCE.THEORY&METHODS 1.676 0.744 13.5 0.00099 0.43 1695 56 0.482 28.19 2.03 240 Q2 1.805 43 Q2 44.296 65.9 Q3 Q3 0.64 Q2 65.909 2020 Soubory v repozitáři: plajner-0519831 Accepted Author Publishing Agreement.pdf 85075330445 SCOPUS PUBMED 000497538000001 WOS PDF 2.8 MB cav_un_epca*0256794 International Journal of General Systems 0308-1079 1563-5104 Roč. 49 č. 1 2020 88 111 Taylor & Francis