Many thanks for this, Jas. I was successfully able to use the revised
version of multinomRob, and it satisfies exactly the needs I was looking
for.
Thanks once again.
Best,
Roger
Jasjeet Singh Sekhon wrote:
As we noted earlier and as is clearly stated in the docs, multinomRob
is
Walter Mebane wrote:
Roger,
Error in if (logliklambda loglik) bvec - blambda :
missing value where TRUE/FALSE needed
In addition: Warning message:
NaNs produced in: sqrt(sigma2GN)
That message comes from the Newton algorithm (defined in source file
multinomMLE.R). It would
As we noted earlier and as is clearly stated in the docs, multinomRob
is estimating an OVERDISPERSED multinomial model. And in your models
here the overdispersion parameter is not identified; you need more
observations. Walter pointed out using the print.level trick to get
the coefs for the
Hi again Jasjeet, Walter,
I have a further question about an error message I get when running
multinomRob. I am simulating a dataset where I look at the effect of
making a previous categorical choice on the probability of making the
same choice later on. Given the following code:
n - 20
Roger,
Error in if (logliklambda loglik) bvec - blambda :
missing value where TRUE/FALSE needed
In addition: Warning message:
NaNs produced in: sqrt(sigma2GN)
That message comes from the Newton algorithm (defined in source file
multinomMLE.R). It would be better if we
Hi Roger,
Yes, multinomRob can handle equality constraints of this type---see
the 'equality' option. But the function assumes that the outcomes are
multinomial counts and it estimates overdispersed multinomial logistic
models via MLE, a robust redescending-M estimator, and LQD which is
another
By default, with print.level=0 or greater, the multinomRob program
prints the maximum likelihood estimates with conventional standard
errors before going on to compute the robust estimates.
Walter Mebane
Jasjeet Singh Sekhon writes:
Hi Roger,
Yes, multinomRob can handle equality
Many thanks for pointing this out to me!
I'm still a bit confused, however, as to how to use multinomRob. For
example I tried to translate the following example using nnet:
x1 - c(1,1,1,1,0,0,0,0,0,0,0,0)
x2 - c(0,0,0,0,1,1,1,1,0,0,0,0)
y - factor(c(a,b,b,c,a,b,c,c,a,a,b,c))
library(nnet)
d -
Roger,
summary(multinomRob(list(y1 ~ x1 + x2,y2 ~ x1 + x2, y3 ~ 0),data=d,
print.level=1))
Walter Mebane
Roger Levy writes:
Many thanks for pointing this out to me!
I'm still a bit confused, however, as to how to use multinomRob. For
example I tried to translate the following
Hi Roger,
Walter's command is correct. To match the exact normalization used by
nnet's multinom(), however, you would need to make the coefficients
zero for the first class (i.e., y1) and not the last (i.e., y3).
mr - multinomRob(list(y2 ~ x1 + x2, y3 ~ x1 + x2, y1~0),data=d,
print.level=1)
I'm interested in doing multinomial logistic regression with equality
constraints on some of the parameter values. For example, with
categorical outcomes Y_1 (baseline), Y_2, and Y_3, and covariates X_1
and X_2, I might want to impose the equality constraint that
\beta_{2,1} = \beta_{3,2}
11 matches
Mail list logo
