R: Calculate point estimate of posterior parameters and create probability distribution

postdist {deal}

R Documentation

Calculate point estimate of posterior parameters and create probability distribution

Description

After finding (learning) posterior parameters, the mean (or mode) of the distributions are found and are used to update the local probability distributions

Usage

postdist(nw) 
postdist.node(nd,nw,vtype = "mode")

Arguments

`nw`	an object of class `network`.
`nd`	an object of class `node`.
`vtype`	a string, taking values "mode" or "mean" of the variance distribution.

Details

In the learning procedure, the postdist.node procedure is called to ensure that the prob property for each node is updated. The prob property is first set in jointprior and is then updated using postdist to update them in the learning procedures. The prob property is used by savenet for exporting a network to the .net format.

Value

an object of class network with updated prob properties for each node.

Author(s)

Susanne Gammelgaard Bøttcher alma@math.auc.dk,
Claus Dethlefsen dethlef@math.auc.dk.

References

Further information about deal can be found at:
http://www.math.auc.dk/novo/deal.