Penaksiran Parameter Generalized Linear Mixed Model Menggunakan Maximum Likelihood

Abstract

The Generalized Linear Mixed Model (GLMM) is a model that combines characteristics from the General Linear Model (GLM) and the Linear Mixed Model (LMM). In this model, the response variable can be derived from an exponential distribution, and it incorporates both fixed and random effects. This study aims to explore parameter estimation methods in GLMM using maximum likelihood. This method is employed to estimate parameters in GLMM and comprehend its application in the analysis of complex data. Adult mortality data in Asia is utilized to test the effectiveness of this method in evaluating the GLMM model. Four predictor variables are investigated: alcohol consumption, measles cases, polio cases, and educational duration. From the research results, the obtained model is lnMit = 6.362151 + 0.059990X1 + 0.000003X2 + 0.001484X3+ … + B44D44 + 0.05980Z. The estimation results indicate that maximum likelihood can provide parameter estimates for GLMM. This method also has advantages in handling data with mixed structures, such as repeated and hierarchical data.