| posteriorDiscard.CatDirichlet | R Documentation |
Contrary to posterior(), this function will update alpha by removing the information of observed samples x for the model structure:
pi|alpha \sim Dir(alpha)
x|pi \sim Categorical(pi)
Where Dir() is the Dirichlet distribution, Categorical() is the Categorical distribution. See ?dDir and dCategorical for the definitions of these distribution.
The model structure and prior parameters are stored in a "CatDirichlet" object, the prior parameters in this object will be updated after running this function.
## S3 method for class 'CatDirichlet' posteriorDiscard(obj, ss, w = NULL, ...)
obj |
A "CatDirichlet" object. |
ss |
Sufficient statistics of x. In Categorical-Dirichlet case the sufficient statistic of sample x can be either x itself, of an "ssCat" object generated by the function sufficientStatistics.CatDirichlet(). |
w |
Sample weights,default NULL. |
... |
Additional arguments to be passed to other inherited types. |
None. the prior parameters stored in "obj" will be updated with the information in "ss".
Murphy, Kevin P. Machine learning: a probabilistic perspective. MIT press, 2012.
CatDirichlet,posterior.CatDirichlet
obj <- CatDirichlet(gamma=list(alpha=rep(1,26),uniqueLabels = letters)) x <- sample(letters,size = 20,replace = TRUE) w <- runif(20) posterior(obj=obj,ss=x) obj posteriorDiscard(obj=obj,ss=x) obj ## weighted sample posterior(obj=obj,ss=x,w=w) obj posteriorDiscard(obj=obj,ss=x,w=w) obj