Computational approach for Bayesian inference

In principal, Bayesian inference is easily implemented based on the posterior distribution of parameters π(θ|Xn) conditional on observed data Xn and prior distribution π(θ). However, in most of practical situations, we don’t have the joint posterior distribution of θ in analytical form. In such a case, we can employ a simulation based approach. This section first introduces the concept of Monte Carlo integration. Then computational approaches for Bayesian inference will be explained.