Computational possibilities and properties of robustified version of the mixed Least Squares - Total Least Squares estimator

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F68407700%3A21340%2F12%3A00199819" target="_blank" >RIV/68407700:21340/12:00199819 - isvavai.cz</a>

Výsledek na webu

—

DOI - Digital Object Identifier

—

Jazyk výsledku

angličtina

Název v původním jazyce

Computational possibilities and properties of robustified version of the mixed Least Squares - Total Least Squares estimator

Popis výsledku v původním jazyce

Classical robust regression estimators, such as Least Trimmed Squares (LTS), are not consistent when both independent and some dependent variables are considered to be measured with a random error. One way how to cope with this problem is to use the robustified version of Mixed Least Squares - Total Least Squares (LS-TLS). Mixed Least Trimmed Squares - Total Least Trimmed Squares (LTS-TLTS) based on trimming and mixed Least Weighted Squares - Total Least Weighted Squares (LWS-TLWS) based on the idea ofdownweighting the influential points, are proposed. The existence and uniqueness of the solution, breakdown point, consistency and another properties of these estimators are discussed. Different approaches of calculation, such as Branch-and-Bound algorithm, elemental concentration algorithm and simulated annealing, are described and their performances are shown on sets of benchmark instances.

Název v anglickém jazyce

Computational possibilities and properties of robustified version of the mixed Least Squares - Total Least Squares estimator

Popis výsledku anglicky

Classical robust regression estimators, such as Least Trimmed Squares (LTS), are not consistent when both independent and some dependent variables are considered to be measured with a random error. One way how to cope with this problem is to use the robustified version of Mixed Least Squares - Total Least Squares (LS-TLS). Mixed Least Trimmed Squares - Total Least Trimmed Squares (LTS-TLTS) based on trimming and mixed Least Weighted Squares - Total Least Weighted Squares (LWS-TLWS) based on the idea ofdownweighting the influential points, are proposed. The existence and uniqueness of the solution, breakdown point, consistency and another properties of these estimators are discussed. Different approaches of calculation, such as Branch-and-Bound algorithm, elemental concentration algorithm and simulated annealing, are described and their performances are shown on sets of benchmark instances.

Druh

O - Ostatní výsledky

CEP obor

BA - Obecná matematika

OECD FORD obor

—

Projekt

—

Návaznosti

S - Specificky vyzkum na vysokych skolach

Rok uplatnění

2012

Kód důvěrnosti údajů

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