PEMODELAN METODE SEEMINGLY UNRELATED REGRESSION (SUR) DAN REGRESI MULTIVARIAT (Aplikasi pada Faktor Yang Mempengaruhi Indeks Risiko Stroke Infark)

ENDANG YUSWATININGSIH, 101041039 (2012) PEMODELAN METODE SEEMINGLY UNRELATED REGRESSION (SUR) DAN REGRESI MULTIVARIAT (Aplikasi pada Faktor Yang Mempengaruhi Indeks Risiko Stroke Infark). Thesis thesis, UNIVERSITAS AIRLANGGA.

[img]
Preview
Text (ABSTRAK)
gdlhub-gdl-s2-2013-yuswatinin-23110-8.abstr-t.pdf

Download (140kB) | Preview
[img] Text (FULL TEXT)
gdlhub-gdl-s2-2013-yuswatinin-23110-15full.pdf
Restricted to Registered users only

Download (400kB) | Request a copy
Official URL: http://lib.unair.ac.id

Abstract

Seemingly Unrelated Regression (SUR) views several equations as a system which generally generates correlation between errors. Due to this correlation, parameter estimation isn’t very efficient. SUR Model tries to streamline parameter estimation in the equations in the form of systems. Analysis results showed that parameter estimation in multivariate regression gave different results with SUR Method. This happened because multivariate regression only used concurrent test so that every predictor variable, systolic blood pressure, diastolic blood pressure, blood sugar during, blood sugar when fasting, blood sugar 2 hours post prandial, HBA1c, total cholesterol and HDL, affected the response variable of myocardial infarction risk index type A and type B. SUR method used concurrent test and partial test, so they could be used to see predictor variables which really affected response variables which were: myocardial infarction risk index type A was effected by systolic blood pressure, diastolic blood pressure, blood sugar during, blood sugar when fasting, blood sugar 2 hours post prandial, HBA1c and total cholesterol, while predictor variables which affected myocardial infarction risk index type B were blood sugar during, blood sugar when fasting, blood sugar 2 hours post prandial and total cholesterol. Modeling with SUR method would generate accurate modeling in which there were the best models for every myocardial infarction risk indexes so the models depended on the significance of each predictor variables and generated a few models in accordance with the dependent variables, compared with global modeling which only generated one model.

Item Type: Thesis (Thesis)
Additional Information: KKC KK TKM 07 - 12 Yus p (FULL TEXT TIDAK LENGKAP)
Uncontrolled Keywords: Multivariate Regression, Seemingly Unrelated Regression (SUR)
Subjects: R Medicine > RC Internal medicine > RC109-216 Infectious and parasitic diseases
Divisions: 10. Fakultas Kesehatan Masyarakat > Magister Ilmu Kesehatan Masyarakat
Creators:
CreatorsEmail
ENDANG YUSWATININGSIH, 101041039UNSPECIFIED
Contributors:
ContributionNameEmail
ContributorSoenarnatalia, Dr., Ir., M.KesUNSPECIFIED
ContributorArief Wibowo, Dr., dr., MSUNSPECIFIED
Depositing User: Nn Dhani Karolyn Putri
Date Deposited: 2016
Last Modified: 19 Oct 2016 03:14
URI: http://repository.unair.ac.id/id/eprint/36800
Sosial Share:

Actions (login required)

View Item View Item