How to Test the Parametric Fit of the Complementary Log Log Model to the Data?

Authors : Ozge AKKUS

Pages : 665-675

View : 19 | Download : 9

Publication Date : 2012-07-19

Article Type : Other Papers

Abstract :In this study, commands written on the basis of Uniform Confidence Bands insert ignore into journalissuearticles values(UCB); in the new windows version of the XploRe package for testing the validity of the link function assumed for binary parametric complementary log log model were introduced. Models including only continuous and both continuous and discrete insert ignore into journalissuearticles values(mixed); explanatory variables cases were discussed, separately. I intended here to present researchers an easier way for testing the accuracy of the assumed parametric complementary log log model for the data that could not be tested by the existing standard statistical packages. The applicability of the commands was shown over an artificial and a real data set on gastric cancer.              Key Words :   Complementary log log model; model validity test; binary dependent variable; semi-parametric     approach; uniform confidence band.  
Keywords : Complementary log log model, model validity test, binary dependent variable, semi parametric