multnom_pred {kDGLM}R Documentation

multnom_pred

Description

Calculate the values for the predictive distribution given the values of the parameter of the conjugated distribution of the linear predictor. The data is assumed to have Multinomial distribution with known number of trial N and the probability vector having distribution Dirichlet with parameters alpha_i. In this scenario, the marginal distribution of the data is Dirichlet-Multinomial with parameters N and alpha_i.

Usage

multnom_pred(conj.param, outcome, parms = list(), pred.cred = 0.95)

Arguments

conj.param

List or data.frame: The parameters of the conjugated distributions of the linear predictor.

outcome

Vector or matrix: The observed values at the current time. The value passed is used to compute N.

parms

List (optional): A list of extra parameters for the model. Not used in this function.

pred.cred

Numeric: the desired credibility for the credibility interval.

Value

A list containing the following values:

See Also

Other auxiliary functions for a Multinomial outcome: convert_Multinom_Normal(), convert_Normal_Multinom(), update_Multinom()


[Package kDGLM version 1.2.0 Index]