20 Mar
2007
20 Mar
'07
6:43 p.m.
There has been some confusion in the lab over what exactly this line of code
does and means:
beta <- par[1:ncol(X)]
Can anyone please explain clearly and in writing ("so I can convince myself
of it later") what this is doing, and why it makes my functions work?
Thanks,
Jill
20 Mar
20 Mar
6:57 p.m.
Jill,
Though it is unnecessary for the probit and logit models used in this and
last week?s problem sets, that command is appropriate when the ?par?
returned by optim() contains estimates of ancillary parameters. ?Ancillary
parameters? are things that have to be estimated to fit a model to the data,
but that are not interesting in and of themselves. The command below
assigns to beta only those estimated parameters which are ?interesting?,
i.e., the estimated coefficients on each of your independent variables.
Andrew
On 3/20/07 6:43 PM, "Jill Goldenziel" <jgolden at fas.harvard.edu> wrote:
There has been some confusion in the lab over what
exactly this line of code
does and means:
beta <- par[1:ncol(X)]
Can anyone please explain clearly and in writing ("so I can convince myself of
it later") what this is doing, and why it makes my functions work?
Thanks,
Jill
_______________________________________________
gov2001-l mailing list
gov2001-l at lists.fas.harvard.edu
http://lists.fas.harvard.edu/mailman/listinfo/gov2001-l
21 Mar
21 Mar
5:15 a.m.
Hi,
I?m not sure what separates the betas from the sigma^2 is ?interestingness?.
As Gary suggested, one may be interested in modelling the variance as a
parameter which depends upon a matrix of explanatory variables Z (where Z
could even be X).
The probit has a ?hidden? ancillary parameter because it assumes unit
variance. (As based on the stylised normal)
From: gov2001-l-bounces at lists.fas.harvard.edu
[mailto:gov2001-l-bounces at lists.fas.harvard.edu] On Behalf Of Andrew Coe
Sent: 20 March 2007 22:57
To: gov2001-l at lists.fas.harvard.edu
Subject: Re: [gov2001-l] beta <- par[1:ncol(X)]
Jill,
Though it is unnecessary for the probit and logit models used in this and
last week?s problem sets, that command is appropriate when the ?par?
returned by optim() contains estimates of ancillary parameters. ?Ancillary
parameters? are things that have to be estimated to fit a model to the data,
but that are not interesting in and of themselves. The command below
assigns to beta only those estimated parameters which are ?interesting?,
i.e., the estimated coefficients on each of your independent variables.
Andrew
On 3/20/07 6:43 PM, "Jill Goldenziel" <jgolden at fas.harvard.edu> wrote:
There has been some confusion in the lab over what exactly this line of code
does and means:
beta <- par[1:ncol(X)]
Can anyone please explain clearly and in writing ("so I can convince myself
of it later") what this is doing, and why it makes my functions work?
Thanks,
Jill
_____
_______________________________________________
gov2001-l mailing list
gov2001-l at lists.fas.harvard.edu
HYPERLINK
"http://lists.fas.harvard.edu/mailman/listinfo/gov2001-l"http://lists.fas.ha
rvard.edu/mailman/listinfo/gov2001-l
--
No virus found in this incoming message.
Checked by AVG Free Edition.
Version: 7.5.446 / Virus Database: 268.18.15/728 - Release Date: 20/03/2007
08:07
--
No virus found in this outgoing message.
Checked by AVG Free Edition.
Version: 7.5.446 / Virus Database: 268.18.15/728 - Release Date: 20/03/2007
08:07
20 Mar
20 Mar
7:10 p.m.
Basically you are taking the ncol(X) values of the parameter vector and
setting them as your betas.
One way to see this is to compare OLS and logit/probit.
Under OLS, your parameters that you want to find (the unknown stuff) are
your betas and your sigma^2. Your parameter vector that you enter into you
function is therefore a vector that stacks your beta guesses with your guess
of sigma^2. The command you have then takes the first ncol(X) values, which
are your betas, and sets them as your betas. You basically extract the
betas from your vector, leaving the sigma^2 behind.
In logit/probit, your parameters are only betas, so you don't actually need
to do that. Since the betas are your only parameters
beta <- par[1:ncol(X)]
is equivalent to
beta <- par
-Patrick
On 3/20/07, Jill Goldenziel <jgolden at fas.harvard.edu> wrote:
There has been some confusion in the lab over what exactly this line of
code does and means:
beta <- par[1:ncol(X)]
Can anyone please explain clearly and in writing ("so I can convince
myself of it later") what this is doing, and why it makes my functions work?
Thanks,
Jill
_______________________________________________
gov2001-l mailing list
gov2001-l at lists.fas.harvard.edu
http://lists.fas.harvard.edu/mailman/listinfo/gov2001-l
6268
days inactive
6269
days old
gov2001@lists.gking.harvard.edu
3 comments
4 participants
participants (4)
-
acoe＠fas.harvard.edu -
jgolden＠fas.harvard.edu -
jm540＠york.ac.uk -
plam＠fas.harvard.edu