Model Marginal Data Longitudinal Biner menggunakan Rantai Markov

Longitudinal data is collected repeatedly over time. Longitudinal data usually have correlation in a form of serial correlation within the subjects. Obviously, if Yit represents the i-th observation subject to time-t, the subject I will take back the response Yit accordingly. Observation result is obtained from the same subject, thus result of these repeated responses are correlated. Analysis of longitudinal data where the response variable is binary is considered as likelihood inference, which requires a complete specification of the stochastic model for the individual. The Markov chain is applied in this study which is a binary-stochastic mechanism to describe the policy-marginal model in longitudinal data with random effects observed. In this thesis, the analysis of binary longitudinal data marginal models using Markov chain with random effects applied to analyze the study in 577 affected infants Acute Respiratory Infection (ARI) in Purworejo. This analysis yields an estimate that is not much different from the logistic regression analysis, but the analysis can be calculated dependencies and variance estimation