llmnp {bayesm}R Documentation

Evaluate Log Likelihood for Multinomial Probit Model

Description

llmnp evaluates the log-likelihood for the multinomial probit model.

Usage

llmnp(beta, Sigma, X, y, r)

Arguments

beta

k x 1 vector of coefficients

Sigma

(p-1) x (p-1) covariance matrix of errors

X

n*(p-1) x k array where X is from differenced system

y

vector of n indicators of multinomial response (1, ..., p)

r

number of draws used in GHK

Details

XX is (p1)nxk(p-1)*n x k matrix. Use createX with DIFF=TRUE to create XX.

Model for each obs: w=Xbeta+ew = Xbeta + e with ee \sim N(0,Sigma)N(0,Sigma).

Censoring mechanism:
if y=j(j<p),wj>max(wj)y=j (j<p), w_j > max(w_{-j}) and wj>0w_j >0
if y=p,w<0y=p, w < 0

To use GHK, we must transform so that these are rectangular regions e.g. if y=1,w1>0y=1, w_1 > 0 and w1w1>0w_1 - w_{-1} > 0.

Define AjA_j such that if j=1,,p1j=1,\ldots,p-1 then Ajw=Ajmu+Aje>0A_jw = A_jmu + A_je > 0 is equivalent to y=jy=j. Thus, if y=jy=j, we have Aje>AjmuA_je > -A_jmu. Lower truncation is Ajmu-A_jmu and cov=AjSigmat(Aj)cov = A_jSigmat(A_j). For j=pj=p, e<mue < - mu.

Value

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

Warning

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

Author(s)

Peter Rossi, Anderson School, UCLA, perossichi@gmail.com.

References

For further discussion, see Chapters 2 and 4, Bayesian Statistics and Marketing by Rossi, Allenby, and McCulloch.

See Also

createX, rmnpGibbs

Examples

## Not run: ll=llmnp(beta,Sigma,X,y,r)

[Package bayesm version 3.1-6 Index]