Publication date: 07/08/2024

Posterior Estimates Report

The Posterior Estimates section of the Bayesian Estimates report in the Repeated Measures Degradation platform contains a table of population parameters and a table of unit parameters. Population parameters refer to the fixed parameters in the model and parameters that describe random parameters; unit parameters refer to the realizations of the population parameters. These tables are summaries of the posterior samples generated by the MCMC procedure. You can save the posterior samples to a data table using the options to the right of the Population Parameters table. For more information about the data export options, see Bayesian Estimates Report Options.

Tip: The posterior data table that is created with the export options to the right of the Population Parameters table contains table scripts that can be used for diagnosing convergence of the posterior distribution. It is important to evaluate the convergence diagnostics in three areas: stationarity, low to moderate autocorrelation, and no strange patterns in the scatterplot. Generally, any shape other than a random or ellipsoid-shaped cloud of points is an indication of possible non-convergence.

The Posterior Estimates red triangle menu contains an option to change the credible level that is used for the credible intervals in both tables in the Posterior Estimates section.

Want more information? Have questions? Get answers in the JMP User Community (community.jmp.com).