Least Squares Method With Equality Constraints and Polynomial Approximation of Lorenz Curve

Result code in IS VaVaI
<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F62690094%3A18450%2F19%3A50015901" target="_blank" >RIV/62690094:18450/19:50015901 - isvavai.cz</a>
Result on the web
<a href="https://mme2019.ef.jcu.cz/files/conference_proceedings.pdf" target="_blank" >https://mme2019.ef.jcu.cz/files/conference_proceedings.pdf</a>
DOI - Digital Object Identifier
—

Result language
angličtina
Original language name
Least Squares Method With Equality Constraints and Polynomial Approximation of Lorenz Curve
Original language description
The least squares method is frequently and successfully used in the solution of various different econometric problems. No restrictions on data sets are usually considered. The only widely known problem with a constraint is the one where a linear model heading through the origin is solved. Nevertheless, we can encounter approximation problems having more data restrictions. For instance, the Lorenz curve which is a curve heading through two given points. In this case, it is useful to apply a least squares method subject to constraints. In this paper, two possible solutions of problems with natural data restrictions are examined. First, it is showed that the constrained problem with two boundary values can be transformed into the classical least squares problem and a special form of the normal equation is derived. A more general problem is then introduced and the Lagrange multiplier method is used to develop a different form of the normal equation. Finally, a polynomial approximation of the Lorenz curve applied to the Czech Republic income data is introduced.
Czech name
—
Czech description
—

Publication year
2019
Confidentiality
S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Article name in the collection
Conference Proceedings, 37th International Conference on Mathematical Methods in Economics 2019
ISBN
978-80-7394-760-6
ISSN
—
e-ISSN
—
Number of pages
6
Pages from-to
332-337
Publisher name
Jihočeská univerzita
Place of publication
České Budějovice
Event location
České Budějovice
Event date
Sep 11, 2019
Type of event by nationality
WRD - Celosvětová akce
UT code for WoS article
000507570400055

Similar results(10)