Kurniasari, Dian and Warsono, Warsono and Widiarti, Widiarti (2012) MODEL BERPANGKAT TIDAK PENUH PADA DATA SPASIAL DENGAN METODE DEKOMPOSISI SPEKTRAL. In: SEMINAR DAN RAPAT TAHUNAN (SEMIRATA) BIDANG ILMU MIPA 2012 BKS PTN BARAT, 11-12 Mei 2012, Universitas Negeri Medan, Sumatera Utara.

[img]
Preview
Text
SEMIRATA BKS-PTN MEDAN 2012_DIAN KURNIASARI.pdf

Download (3MB) | Preview

Abstract

Not full rank model is a linear model where matrix data is not full rank.One case of this model in spatial data. Spatial Data is represents world phenomenon that has coordinate reference such as map, plane's or satelite's capture or the result of these representation. The observation relationship on data spatial causes the sample data observation tend to influence and independence each other. Which cause autocorrelation and heteroscedasticity. To overcome properties, we can be use Generalized Least Square (GLS) method. The estimation of \Beta can be derived by minimizing sum square of error that has transformation by non singular matrix P, which is got from decomposition of variance of error singular matrix using spectral decomposition. In this research, the implementation of algorithm for not full rank model with GLS method done by using data simulation using SAS. Variance of error which is resulted is \Delta= I-X(X^TX)^-1X^T. The result of simulation study shown that non full rank odel can be used to overcome the problem of autocorrelation, heteroscedaticity on spatial data.

Item Type: Conference or Workshop Item (Paper)
Subjects: Q Science > QA Mathematics
Divisions: Fakultas Matematika dan Ilmu Pengetahuan Alam (FMIPA) > Prodi Matematika
Depositing User: DIAN KURNIASARI
Date Deposited: 12 Jun 2017 02:04
Last Modified: 12 Jun 2017 02:04
URI: http://repository.lppm.unila.ac.id/id/eprint/3469

Actions (login required)

View Item View Item