|
Výpočetní aspekty návrhu regulátoru a vyčíslení kvality
Novák, Miroslav
This work contributes to the activity in the Department of Adaptive Systems in Institute of Information Theory and Automation, Academy of Sciences of the Czech Republic to develop a complete design algorithm for advanced controllers such as the LQG one and put them through to real applications. The task of controller tuning is to transform the user specified requirements into the values of the tuning parameters. The system knowledge is incomplete. The Bayesian estimation delivers the parameters not as known numbers but as their probability density function. The important contribution of this work is extending the tuning to the multiple input multiple output (MIMO) controllers, where multiple constraints on particular quantities are considered simultaneously.
|
| |
|
Modifikovaná Raova vzdálenost
Fabián, Zdeněk
V příspěvku je pod názvem Johnsonova skórová funkce představena vlivová funkce pravděpodobnostního rozdělení. V případě parametrického rozdělení ji lze chápat jako inferenční funkci, kterou lze použít ke konstrukci charakteristik rozdělení i datových souborů a k modifikaci Raovy vzdálenosti pro testování hypotéz.
|
|
Johnsonův bod a Johnsonova disperze
Fabián, Zdeněk
New measures of central tendency and dispersion of continuous probability distributions were introduced. In this paper we show that the theory offer a simple and suitable description of heavy-tailed distributions, i.e. the distributions which may not have the mean and/or variance.
|
|
Robustnost mediánového odhadu v Bernoulliově logistické regresi
Hobza, Tomáš ; Pardo, L.
In the paper the median estimator of the logistic regression parameters employing smoothed data in the discrete case is considered. Sensitivity of this estimator to contaminations of the logistic regression data is studied by simulations and compared with the sensitivity of some robust estimators previously introduced to logistic regression.
|
| |
| |
|
Divergence of proobabilty ditributions and statistical information
Vajda, Igor
A recently established generalized Taylor formula is used to prove in a new much simpler manner the basic properties of f-divergences of probability measures. In a new simple way is proved also the relation of these divergences to the statistical information defined as a measure of decrease of the Bayes risk.
|
| |
| |