ESTIMASI BAYESIAN PADA MODEL PERSAMAAN STRUKTURAL DENGAN VARIABEL KATEGORIK TERURUT; BAYESIAN ESTIMATION ON STRUCTURAL EQUATION MODELING WITH ORDERED CATEGORICAL VARIABLE

RINI YUNITA, Subanar

2013 | Disertasi | PROGRAM STUDI S2 MATEMATIKA

This thesis describes the parameter estimation of structural equation models with ordered categorical variables using Bayesian methods. The basic assumption of SEM-based covarian is interval scale data and meet the assumptions of normality. Ordinal categorical data can be used as normally distributed continuous data by searching for each threshold parameter data. Bayes methods analyze sample data with prior information into account, in order to minimize the error rate. The estimation process performed numerically using a Monte Carlo method, namely Gibbs Sampling and Metropolis Hasting

Kata Kunci : Model Persamaan Struktural; Metode Bayes; Data Kategorik; Threshold; Gibbs Sampling; Metropolis Hasting