Approximate Bayesian Computation And Its Application

In this article we introduced the approximate Bayesian computation(ABC)and use this method to infer parameters in three statistic models.The foundation of ABC is Bayesian computation.With the provided model and prior distribution we can infer posterior distribution from the observed data.Different from other Bayesian computation method,ABC doesn't require a strict likelihood function and it is very easy to be applied in most mathematic models.ABC can also be combined with other sampling method.This article use ABC to infer parameters in binomial model,hierarchical binomial model and time series models,it helped us to get the posterior distribution correctly and efficiently.Compare to the other parameter estimation method,ABC tends to be more convenient and efficient.