Steps for a well Bayesian Statistical Inference

Steps for a well Bayesian Statistical Inference

1. The accurate formulation of the data likelihood

2. The choice of the prior, which incorporates all other knowledge tha might exist, but it is not used when computing the likelihood

   • Determination of the posterior pdf, using Bayes' theorem. In practice, this step can be computationally intensive for complex multimentional problems
     • The search for a best model parameters M, which maximize ,yielding the maximum a posteriori (MAP) estimate
       • We can simulate the samples from the posterior from the true underlying distribution of data and estimators of models parameters based on data.

Images and files in the map