UMM Institutional Repository

MENENTUKAN INTERVAL KEPERCAYAAN PARAMETER MODEL REGRESI NONLINIER POWER DENGAN METODE OLS (Ordinary Least Square) DAN GLS (Generalized Least Square)

PUSPITASARI, FIKA (2014) MENENTUKAN INTERVAL KEPERCAYAAN PARAMETER MODEL REGRESI NONLINIER POWER DENGAN METODE OLS (Ordinary Least Square) DAN GLS (Generalized Least Square). Other thesis, University of Muhammadiyah Malang.

Full text not available from this repository.

Abstract

Parameter estimation is the statement concerning an unknown population parameter based on the population of sample in this case it is random sample taken from the population concerned so that with this prediction, the state parameter of the population is knowable. The best expectation are expected to be used to predict the value of Y properly. Parameter estimation is divided into two these are point estimate and interval estimate. In practice, the point estimate is comprised of one single value doesn't give you anidea of how the distance or the difference between the value of the expectation to the true value.The most commonly used methods in parameter estimation is a method of OLS (Ordinary Least Square).To use this method, the error of the regression model disturber must fulfill the assumptions of classical linear regression model (Classical Linear Regression Model). By using OLS method obtained linear expectation, unbiased and best (the minimum variance).However, in practice, often going on diversion model assumptions of classical linear regression, one of which is heteroskedastisitas, where the variance assumption is no longer constant.When such distortion occurs, then the OLS method is no longer appropriate to use in parameter estimation. There is a method which is the development of methods of OLS method, that is GLS (Generalized Least Square). The GLS method can be used in the data homoskedastisitas and is also used to address heteroskedastisitas. After the regression model parameters obtained expectation of nonlinear power, then the confidence interval can also be known. The confidence Interval is used as something of parameter estimates, because almost never found exactly the same statistical value with its parameters. By using the method of OLS and GLS method with confidence interval obtained GLS is shorter than OLS method, it means that the confidence interval method with GLS more ideal than the OLS method.

Item Type: Thesis (Other)
Subjects: Q Science > QA Mathematics
Divisions: Faculty of Teacher Training and Education > Department of Mathematics and Computing
Depositing User: Gusti Vani Putri Cahya
Date Deposited: 27 Oct 2008 05:38
Last Modified: 27 Oct 2008 05:38
URI : http://eprints.umm.ac.id/id/eprint/15954

Actions (login required)

View Item View Item
UMM Official

© 2008 UMM Library. All Rights Reserved.
Jl. Raya Tlogomas No 246 Malang East Java Indonesia - Phone +62341464318 ext. 150, 151 - Fax +62341464101
E-Mail : infopus[at]umm.ac.id - Website : http://lib.umm.ac.id - Online Catalog : http://laser.umm.ac.id - Repository : http://eprints.umm.ac.id

Web Analytics

UMM Institutional Repository is powered by :
EPrints Logo