To learn more about EpiX Analytics' work, please visit our modeling applications, white papers, and training schedule.
Page tree
Skip to end of metadata
Go to start of metadata

 

 

Prior distributions

 

The prior distributions are the description of one's state of knowledge about the parameter in question prior to observation of the data. Determination of the prior distribution is the primary focus for criticism of Bayesian inference and one needs to be quite sure of the effects of choosing one particular prior over another. This section describes three different types of prior distributions:

 

Uninformed priors – describing that you have no prior knowledge

Conjugate priors – a parametric distribution that can be easily updated

Subjective priors – a distribution constructed from an expert's opinion

Improper priors – a prior distribution that does not normalize to unity

Informed priors - a description of the level of knowledge you have

 

 

 


  • No labels