Open AccessPERBANDINGAN MODEL REGRESI KEGAGALAN PROPORSIONAL DARI COX MENGGUNAKAN METODE EFRON DAN EXACT
Author(s) -
Asri Lutfia Silmi,
Sudarno Sudarno,
Puspita Kartikasari
Publication year - 2020
Publication title -
jurnal gaussian : jurnal statistika undip
Language(s) - English
Resource type - Journals
ISSN - 2339-2541
DOI - 10.14710/j.gauss.v9i4.29008
Subject(s) - proportional hazards model , survival analysis , statistics , hazard ratio , hazard , accelerated failure time model , mathematics , econometrics , event (particle physics) , regression analysis , confidence interval , physics , biology , ecology , quantum mechanics
Cox proportional hazard regression analysis is one of statistical methods that is often used in survival analysis to determine the effect of independent variables on the dependent variable in the form of survival time. Survival time starts from the beginning of the study until the event occurs or has reached the end of the study. The Cox proportional hazard regression model does not require information about the distribution that underlies the survival time but there is an assumption of proportional hazard that must be met. The purpose of this study is to determine the factors that influence the survival time of coronary heart disease. Ties are often found in survival data, including the survival data used in this study. Ties is an event when there are two or more individuals who experience a failure at the same time or have the same survival time value. The Efron and Exact method approach is used to overcome the presence of ties that can cause problems in the estimation of parameters associated with determining the members of the risk set. The results showed that the variables of diabetes mellitus, family history, and platelets significantly affected the survival time of CHD patients for both methods. The best model obtained is the Exact method because it has smaller AIC value of 383,153 compared to the AIC value of the Efron method of 393,207.