Try removing the percent signs from the row "c": c<- 1+exp(b*a) and adding them in the row "a": a<-(X%*%beta) ________________________________ From: gov2001-l-bounces at lists.fas.harvard.edu on behalf of Keith Schnakenberg Sent: Sat 3/15/2008 18:12 To: gov2001-l at lists.fas.harvard.edu Subject: [gov2001-l] problems with logit log likelihood I'm having problems implementing the log likelihood function. I used the following code: loglogit <- function(par, X=X, y=y){ beta <- par[1:ncol(X)] a <- (X*beta) b <- (1 - (2*y)) c <- 1 + exp(b%*%a) d <- log(c) out <- sum(d) } And when I give it to optim I get very unlikely coefficients and a Hessian of all zeros. I can't seem to figure out what I am overlooking. Does anybody see any obvious problems with this function? If so, I would appreciate a nudge in the right direction. Thanks, Keith _______________________________________________ gov2001-l mailing list gov2001-l at lists.fas.harvard.edu http://lists.fas.harvard.edu/mailman/listinfo/gov2001-l _______________________________________________ gov2001-l mailing list gov2001-l at lists.fas.harvard.edu http://lists.fas.harvard.edu/mailman/listinfo/gov2001-l

That is a good point, but does not actually fix the problem. Also, when I change the percentage signs, optim() tells me something about non-finite finite-difference values. Thanks, Keith _______________________________________________ gov2001-l mailing list gov2001-l at lists.fas.harvard.edu http://lists.fas.harvard.edu/mailman/listinfo/gov2001-l