Uji Park Dan Uji Breusch Pagan Godfrey Dalam Pendeteksian Heteroskedastisitas Pada Analisis Regresi

Main Author: Andriani, Siska
Format: Article info application/pdf eJournal
Bahasa: eng
Terbitan: Raden Intan State Islamic University of Lampung, INDONESIA , 2017
Subjects:
Online Access: http://ejournal.radenintan.ac.id/index.php/al-jabar/article/view/1014
http://ejournal.radenintan.ac.id/index.php/al-jabar/article/view/1014/907
Daftar Isi:
  • Homoskedastisitas is one of the conditions are fulfilled classical assumptions in the regression analysis, if not met this means homoskedastisitas error variance is not constant and is said to occur heteroscedasticity problem. Test Park and Pagan Godfrey Breusch test is a statistical test to detect whether there is a problem of heteroscedasticity in the regression equation. The problem is how to test the results of detection heteroskedastisitas Park and Breusch Pagan Godfrey test, which is more effective test.Based on the results of research and discussion can be concluded that the detection of the three cases of data acquired two pieces of data in the test with both test detected heteroskedasticity problems, while one case is detected by the test Breusch heteroskedastisitas Pagan Godfrey Park but the test was not detected. Values mean square error (MSE) test Breusch Pagan Godfrey smaller than the test Park so it can be said Pagan Godfrey Breusch method used more effectively. Thus, in detecting problems hetereoskedastisitas should use Breusch Pagan Godfrey test because they have better accuracy than tests Park.
  • The purpose of this study was to obtain an overview of the detection of heteroscedasticity through the parkfrey test and pagan breeding tests in the regression analysis, which of the two tests were more effective. Research methods using qualitative research methods. The type of qualitative research method used is a type of development study research method. The development study referred to is to obtain a developmental picture of the two tests, namely the parkfitness test and the godfrey pagan breed test in the detection of heteroscedasticity in regression analysis through cross sectional method. Based on the results of detection studies in three data cases, two data were obtained which were tested by testing detected heteroscedasticity problems. Others there was one case that detected heteroscedasticity with the godfrey pagan breed test, but for the case with an undetectable park test. The results of the mean square error (MSE) obtained through the godfrey pagan breed test were smaller than the park test. It can be concluded that the godfrey pagan breeding test is more effective to use, compared to the park test in detecting heteroscedasticity in regression analysis. This means that in detecting the hetereoskedasticity problem, it is better to use the godfrey pagan breed test because it has better accuracy than the park test.