Some computational aspects of robustifed total least squares

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F68407700%3A21340%2F11%3A00189482" target="_blank" >RIV/68407700:21340/11:00189482 - isvavai.cz</a>

Výsledek na webu

—

DOI - Digital Object Identifier

—

Jazyk výsledku

angličtina

Název v původním jazyce

Some computational aspects of robustifed total least squares

Popis výsledku v původním jazyce

Classical regression estimators, such as the Ordinary Least Squares, are very sensitive to occurrence of outliers and they are not consistent when orthogonality condition is broken and both independent and some dependent variables are considered to be measured with a random error. The method which can cope with this problem is robustified version of the mixed Least Squares - Total Least Squares method, based on the idea of down-weighting the influential points. The existence and the uniqueness of the solution is discussed and di(R)erent approaches of calculation are described. The computational complexity is shown and the exact algorithm based on a branch-and-bound technique that guarantees global optimality is presented. The implementation of the algorithms and theirs application to simulated and real data sets are shown.

Název v anglickém jazyce

Some computational aspects of robustifed total least squares

Popis výsledku anglicky

Classical regression estimators, such as the Ordinary Least Squares, are very sensitive to occurrence of outliers and they are not consistent when orthogonality condition is broken and both independent and some dependent variables are considered to be measured with a random error. The method which can cope with this problem is robustified version of the mixed Least Squares - Total Least Squares method, based on the idea of down-weighting the influential points. The existence and the uniqueness of the solution is discussed and di(R)erent approaches of calculation are described. The computational complexity is shown and the exact algorithm based on a branch-and-bound technique that guarantees global optimality is presented. The implementation of the algorithms and theirs application to simulated and real data sets are shown.

Druh

D - Stať ve sborníku

CEP obor

BA - Obecná matematika

OECD FORD obor

—

Projekt

—

Návaznosti

S - Specificky vyzkum na vysokych skolach

Rok uplatnění

2011

Kód důvěrnosti údajů

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Název statě ve sborníku

SPMS 2011 Stochastic and Physical Monitoring Systems - Proceedings

ISBN

978-80-01-04916-7

ISSN

—

e-ISSN

—

Počet stran výsledku

13

Strana od-do

35-47

Název nakladatele

ČVUT, Fakulta jaderná a fyzikálně inženýrská

Místo vydání

Praha

Místo konání akce

Křižánky

Datum konání akce

27. 6. 2011

Typ akce podle státní příslušnosti

EUR - Evropská akce

Kód UT WoS článku

—