Maximum a posteriori estimation

This article needs additional citations for verification. Please help improve this article by adding citations to reliable sources. Unsourced material may be challenged and removed. Find sources: "Maximum a posteriori estimation" – news · newspapers · books · scholar · JSTOR (September 2011) (Learn how and when to remove this template message)

Part of a series on
Bayesian statistics
Posterior = Likelihood × Prior ÷ Evidence
Background
Bayesian inference Bayesian probability Bayes' theorem Bernstein–von Mises theorem Coherence Cox's theorem Cromwell's rule Principle of indifference Principle of maximum entropy
Model building
Strong prior Conjugate prior Linear regression Empirical Bayes Hierarchical model
Posterior approximation
Markov chain Monte Carlo Laplace's approximation Integrated nested Laplace approximations Variational inference Approximate Bayesian computation
Estimators
Bayesian estimator Credible interval Maximum a posteriori estimation
Evidence approximation
Evidence lower bound Nested sampling
Model evaluation
Bayes factor Model averaging Posterior predictive
Mathematics portal
v t e

In

of maximum likelihood estimation.

Description

Assume that we want to estimate an unobserved population parameter ${\displaystyle \theta }$ on the basis of observations ${\displaystyle x}$. Let ${\displaystyle f}$ be the sampling distribution of ${\displaystyle x}$, so that ${\displaystyle f(x\mid \theta )}$ is the probability of ${\displaystyle x}$ when the underlying population parameter is ${\displaystyle \theta }$. Then the function:

${\displaystyle \theta \mapsto f(x\mid \theta )\!}$

is known as the likelihood function and the estimate:

${\displaystyle {\hat {\theta }}_{\mathrm {MLE} }(x)={\underset {\theta }{\operatorname {arg\,max} }}\ f(x\mid \theta )\!}$

is the maximum likelihood estimate of ${\displaystyle \theta }$.

Now assume that a

${\displaystyle g}$ over ${\displaystyle \theta }$ exists. This allows us to treat ${\displaystyle \theta }$ as a of ${\displaystyle \theta }$ using Bayes' theorem:

${\displaystyle \theta \mapsto f(\theta \mid x)={\frac {f(x\mid \theta )\,g(\theta )}{\displaystyle \int _{\Theta }f(x\mid \vartheta )\,g(\vartheta )\,d\vartheta }}\!}$

where ${\displaystyle g}$ is density function of ${\displaystyle \theta }$, ${\displaystyle \Theta }$ is the domain of ${\displaystyle g}$.

The method of maximum a posteriori estimation then estimates ${\displaystyle \theta }$ as the mode of the posterior distribution of this random variable:

${\displaystyle {\begin{aligned}{\hat {\theta }}_{\mathrm {MAP} }(x)&={\underset {\theta }{\operatorname {arg\,max} }}\ f(\theta \mid x)\\&={\underset {\theta }{\operatorname {arg\,max} }}\ {\frac {f(x\mid \theta )\,g(\theta )}{\displaystyle \int _{\Theta }f(x\mid \vartheta )\,g(\vartheta )\,d\vartheta }}\\&={\underset {\theta }{\operatorname {arg\,max} }}\ f(x\mid \theta )\,g(\theta ).\end{aligned}}\!}$

The denominator of the posterior distribution (so-called marginal likelihood) is always positive and does not depend on ${\displaystyle \theta }$ and therefore plays no role in the optimization. Observe that the MAP estimate of ${\displaystyle \theta }$ coincides with the ML estimate when the prior ${\displaystyle g}$ is uniform (i.e., ${\displaystyle g}$ is a constant function).

When the loss function is of the form

${\displaystyle L(\theta ,a)={\begin{cases}0,&{\text{if }}|a-\theta |<c,\\1,&{\text{otherwise}},\\\end{cases}}}$

as ${\displaystyle c}$ goes to 0, the Bayes estimator approaches the MAP estimator, provided that the distribution of ${\displaystyle \theta }$ is quasi-concave.^[1] But generally a MAP estimator is not a Bayes estimator unless ${\displaystyle \theta }$ is

.

Computation

MAP estimates can be computed in several ways:

1. Analytically, when the mode(s) of the posterior distribution can be given in
   closed form. This is the case when conjugate priors
   are used.
2. Via
   optimization such as the conjugate gradient method or Newton's method. This usually requires first or second derivatives
   , which have to be evaluated analytically or numerically.
3. Via a modification of an
   expectation-maximization algorithm
   . This does not require derivatives of the posterior density.
4. Via a Monte Carlo method using simulated annealing

Limitations

While only mild conditions are required for MAP estimation to be a limiting case of Bayes estimation (under the 0–1 loss function),^[1] it is not very representative of Bayesian methods in general. This is because MAP estimates are point estimates, whereas Bayesian methods are characterized by the use of distributions to summarize data and draw inferences: thus, Bayesian methods tend to report the posterior mean or median instead, together with credible intervals. This is both because these estimators are optimal under squared-error and linear-error loss respectively—which are more representative of typical loss functions—and for a continuous posterior distribution there is no loss function which suggests the MAP is the optimal point estimator. In addition, the posterior distribution may often not have a simple analytic form: in this case, the distribution can be simulated using Markov chain Monte Carlo techniques, while optimization to find its mode(s) may be difficult or impossible.^{[citation needed]}

bimodal distribution

in which the highest mode is uncharacteristic of the majority of the distribution

In many types of models, such as

issues arise). Furthermore, the highest mode may be uncharacteristic of the majority of the posterior.

Finally, unlike ML estimators, the MAP estimate is not invariant under reparameterization. Switching from one parameterization to another involves introducing a Jacobian that impacts on the location of the maximum.^[2]

As an example of the difference between Bayes estimators mentioned above (mean and median estimators) and using a MAP estimate, consider the case where there is a need to classify inputs ${\displaystyle x}$ as either positive or negative (for example, loans as risky or safe). Suppose there are just three possible hypotheses about the correct method of classification ${\displaystyle h_{1}}$, ${\displaystyle h_{2}}$ and ${\displaystyle h_{3}}$ with posteriors 0.4, 0.3 and 0.3 respectively. Suppose given a new instance, ${\displaystyle x}$, ${\displaystyle h_{1}}$ classifies it as positive, whereas the other two classify it as negative. Using the MAP estimate for the correct classifier ${\displaystyle h_{1}}$, ${\displaystyle x}$ is classified as positive, whereas the Bayes estimators would average over all hypotheses and classify ${\displaystyle x}$ as negative.

Example

Suppose that we are given a sequence ${\displaystyle (x_{1},\dots ,x_{n})}$ of

${\displaystyle N(\mu ,\sigma _{v}^{2})}$ random variables and a prior distribution of ${\displaystyle \mu }$ is given by ${\displaystyle N(\mu _{0},\sigma _{m}^{2})}$. We wish to find the MAP estimate of ${\displaystyle \mu }$. Note that the normal distribution is its own analytically.

The function to be maximized is then given by

${\displaystyle f(\mu )f(x\mid \mu )=\pi (\mu )L(\mu )={\frac {1}{{\sqrt {2\pi }}\sigma _{m}}}\exp \left(-{\frac {1}{2}}\left({\frac {\mu -\mu _{0}}{\sigma _{m}}}\right)^{2}\right)\prod _{j=1}^{n}{\frac {1}{{\sqrt {2\pi }}\sigma _{v}}}\exp \left(-{\frac {1}{2}}\left({\frac {x_{j}-\mu }{\sigma _{v}}}\right)^{2}\right),}$

which is equivalent to minimizing the following function of ${\displaystyle \mu }$:

${\displaystyle \sum _{j=1}^{n}\left({\frac {x_{j}-\mu }{\sigma _{v}}}\right)^{2}+\left({\frac {\mu -\mu _{0}}{\sigma _{m}}}\right)^{2}.}$

Thus, we see that the MAP estimator for μ is given by

${\displaystyle {\hat {\mu }}_{\mathrm {MAP} }={\frac {\sigma _{m}^{2}\,n}{\sigma _{m}^{2}\,n+\sigma _{v}^{2}}}\left({\frac {1}{n}}\sum _{j=1}^{n}x_{j}\right)+{\frac {\sigma _{v}^{2}}{\sigma _{m}^{2}\,n+\sigma _{v}^{2}}}\,\mu _{0}={\frac {\sigma _{m}^{2}\left(\sum _{j=1}^{n}x_{j}\right)+\sigma _{v}^{2}\,\mu _{0}}{\sigma _{m}^{2}\,n+\sigma _{v}^{2}}}.}$

which turns out to be a linear interpolation between the prior mean and the sample mean weighted by their respective covariances.

The case of ${\displaystyle \sigma _{m}\to \infty }$ is called a non-informative prior and leads to an ill-defined a priori probability distribution; in this case ${\displaystyle {\hat {\mu }}_{\mathrm {MAP} }\to {\hat {\mu }}_{\mathrm {MLE} }.}$

References

• DeGroot, M. (1970). Optimal Statistical Decisions. McGraw-Hill.
  ISBN 0-07-016242-5
  .
• Sorenson, Harold W. (1980). Parameter Estimation: Principles and Problems. Marcel Dekker.
  ISBN 0-8247-6987-2
  .
• Hald, Anders (2007). "Gauss's Derivation of the Normal Distribution and the Method of Least Squares, 1809". A History of Parametric Statistical Inference from Bernoulli to Fisher, 1713–1935. New York: Springer. pp. 55–61.
  ISBN 978-0-387-46409-1
  .