Regression among parts of compositional data with an economic application

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F61989592%3A15310%2F14%3A33153544" target="_blank" >RIV/61989592:15310/14:33153544 - isvavai.cz</a>

Výsledek na webu

—

DOI - Digital Object Identifier

—

Jazyk výsledku

angličtina

Název v původním jazyce

Regression among parts of compositional data with an economic application

Popis výsledku v původním jazyce

Compositional data follow the Aitchison geometry on the simplex. Therefore, it is not appropriate to apply regression analysis directly to the raw compositional data. For this purpose, isometric logratio (ilr) coordinates were formed to express compositional data in coordinates with respect to an orthonormal basis on the simplex. For regression analysis between the response part and an explanatory subcomposition such coordinates are required that enable assigning coordinates to both the response variable and one of the covariates simultaneously. Moreover, as all coordinates contain measurement errors, the use of ordinary least squares regression for parameter estimation would lead to biased results. Therefore, orthogonal regression is used leading to an intuitive geometric interpretation. In the application part, the relation between different activities of gross value added (that form inherently compositional data) are investigated in ilr coordinates.

Název v anglickém jazyce

Regression among parts of compositional data with an economic application

Popis výsledku anglicky

Compositional data follow the Aitchison geometry on the simplex. Therefore, it is not appropriate to apply regression analysis directly to the raw compositional data. For this purpose, isometric logratio (ilr) coordinates were formed to express compositional data in coordinates with respect to an orthonormal basis on the simplex. For regression analysis between the response part and an explanatory subcomposition such coordinates are required that enable assigning coordinates to both the response variable and one of the covariates simultaneously. Moreover, as all coordinates contain measurement errors, the use of ordinary least squares regression for parameter estimation would lead to biased results. Therefore, orthogonal regression is used leading to an intuitive geometric interpretation. In the application part, the relation between different activities of gross value added (that form inherently compositional data) are investigated in ilr coordinates.

Druh

D - Stať ve sborníku

CEP obor

BB - Aplikovaná statistika, operační výzkum

OECD FORD obor

—

Projekt

—

Návaznosti

S - Specificky vyzkum na vysokych skolach

Rok uplatnění

2014

Kód důvěrnosti údajů

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

Název statě ve sborníku

Mathematical Methods in Economics MME 2014 Conference Proceedings

ISBN

978-80-244-4209-9

ISSN

—

e-ISSN

—

Počet stran výsledku

6

Strana od-do

337-342

Název nakladatele

Univerzita Palackého

Místo vydání

Olomouc

Místo konání akce

Olomouc

Datum konání akce

10. 9. 2014

Typ akce podle státní příslušnosti

WRD - Celosvětová akce

Kód UT WoS článku

—