bayesm (version 3.1-6)

llmnl: Evaluate Log Likelihood for Multinomial Logit Model

Description

llmnl evaluates log-likelihood for the multinomial logit model.

Usage

llmnl(beta, y, X)

Value

Value of log-likelihood (sum of log prob of observed multinomial outcomes).

Arguments

beta

$kx1$ coefficient vector

y

$nx1$ vector of obs on y (1,..., p)

X

$n\ast pxk$ design matrix (use createX to create $X$)

Warning

This routine is a utility routine that does not check the input arguments for proper dimensions and type.

Author

Peter Rossi, Anderson School, UCLA, .

Details

Let ${\mu }_i=X_{i}beta$, then $Pr(y_i=j)=exp({\mu }_{i,j})/\sum _{k}exp({\mu }_{i,k})$.
$X_i$ is the submatrix of $X$ corresponding to the $i$th observation. $X$ has $n\ast p$ rows.

Use createX to create $X$.

References

For further discussion, see Bayesian Statistics and Marketing by Rossi, Allenby, and McCulloch.

See Also

createX, rmnlIndepMetrop

Examples

if (FALSE) ll=llmnl(beta,y,X)