Naslov
The least absolute deviation linear regression: properties and two efficient methods

Autori
Kuzmanović, Ivana ; Sabo, Kristian ; Scitovski, Rudolf ; Vazler, Ivan

Vrsta, podvrsta i kategorija rada
Radovi u zbornicima skupova, cjeloviti rad (in extenso), znanstveni

Izvornik
Aplimat-Journal of Applied Mathematics / Kováčová Monika - Bratislava : Slovak University of Technology Press, 2009, 227-240

Skup
Aplimat

Mjesto i datum
Bratislava, Slovačka, 05.02.2008. - 08.02.2008

Vrsta sudjelovanja
Predavanje

Vrsta recenzije
Međunarodna recenzija

Ključne riječi
least absolute deviation line; least absolute deviation plane; linear regression
(least absolute deviation line; least absolute deviation plane linear regression)

Sažetak
For the given set of data, among which outliers(wild points) are expected, the problem of determining the best Least Absolute Deviations (LAD) linear regression is considered. Particulary, the problem of determinig the best weighted LAD-line and the best LAD-plane is considered and efficient algorithms for solving these problems are given. Algorithms are illustrated by several examples as well compared with other methods known in literature. The proposed methods proved to be sufficently efficient for being considered as giving a solution in real time. Therefore, they are suitable for various applications, as e.g. in robotics.

Izvorni jezik
Engleski

Znanstvena područja
Matematika

POVEZANOST RADA