## Baidu Encyclopedia version

In statistics, MAP is an abbreviation for Maximum a posteriori. The estimation method obtains a point estimate for an unobservable amount based on empirical data. It is closely related to the Fisher method in maximum likelihood estimation, but it uses an increased optimization goal, which incorporates the prior distribution of the estimator. So the maximum a posteriori estimate can be seen as the maximum likelihood estimate for regularization.

## Wikipedia version

In Bayesian statistics, a maximum a posteriori probability (MAP) estimate is an unknown number, ie, a posterior distribution equal to the estimated mode. The MAP can be used to obtain an unobserved point estimate based on empirical data. It is closely related to the Maximum Likelihood (ML) estimation method, but uses an enhanced optimization goal (a priori that is quantified by prior knowledge of related events) containing a priori distribution to exceed the number of estimates. Therefore, the MAP estimate can be considered as a regularization of the ML estimate.