ESTIMASI MODEL CAMPURAN DUA DISTRIBUSI DENGAN ALGORITMA EKSPEKTASI MAKSIMISASI PADA DATA SURVIVAL; (ESTIMATED MIXTURE MODELS OF TWO DISTRIBUTION WITH EXPECTATION MAXIMIZATION ALGORITHM FOR SURVIVAL DATA)

ASTUTI, DENY BUDI, Zulaela

2016 | Skripsi | FMIPA

Mixture model is a combination model of two or more distributions. Distribution in the model mixture can consist of distribution identic with different parameter or non-identic. This model classify data within dataset into groups of data that previously were not defined. In the mixture two distribution models, exponential and Weibull distribution make three different models. This models are mixture models of exponential-exponential, exponential-Weibull, Weibull-Weibull. In the mixture two distribution models will be give weight of the two distributions. The sum of the weights should be one. To look the estimation of exponential Weibull mixture model used method algorithm EM because asumsy missing data. A case study on patient-lung-data by programming R, we concluded that the three models that form a mixture of two distributian coresponding to the dataset lung. Mixture model of Weibull-Weibull is the best model for dataset the lung.

Kata Kunci : survival analysis, model mixture two distribution, algorithm EM.