Drexel dragonThe Math ForumDonate to the Math Forum



Search All of the Math Forum:

Views expressed in these public forums are not endorsed by Drexel University or The Math Forum.


Math Forum » Discussions » sci.math.* » sci.stat.math.independent

Topic: simulation question
Replies: 3   Last Post: Feb 28, 2013 1:06 AM

Advanced Search

Back to Topic List Back to Topic List Jump to Tree View Jump to Tree View   Messages: [ Previous | Next ]
Bijan Mobasseri

Posts: 18
Registered: 12/7/04
Re: simulation question
Posted: Feb 27, 2013 9:21 PM
  Click to see the message monospaced in plain text Plain Text   Click to reply to this topic Reply

On 2013-02-27 04:38:41 +0000, Rich Ulrich said:

> On Tue, 26 Feb 2013 16:59:25 -0800, Mike Hollis <email@domain.com>
> wrote:
>

>> I'm trying to formulate a simulation program in Stata. In particular,
>> I want to simulation the behavior of a regression model with two
>> independent variables under specific conditions. Stata has a function
>> for developing oversavtions from a multivariate normal distribution

>
> - took a second, but I figure "oversavtions" is "observations" -


Yes. Sorry about the typo.
>
>> give a correlation/covariance matrix and vectors of means and standard
>> deviations. But I'd rather formulate the model explicitly--something I
>> haven't done since grad school, many years ago.

>
> You lose me here. What do *you* mean by "formulate ... explicitly"?
> I figure that the first step is "developing observations" which you
> then might do something else to (your "specific conditions").


i can use observtions generated from a multivariate nornal
distirbution, pick a large n and run a regression to find values for
the regression coefficients or I can explicitly formulate and solve for
the betas using covariance algebra. It's the later that I'm interested
in.

>
>>
>> Assume I have a standardize regression model with a known correlation
>> matrix. I can set one beta and I'd like to derive the expression for

>
> What do you mean by "set one beta"? Fixing your correlation matrix
> sort of sets both ....


Yes, if I generate observations from a multivariate normal dist. as
described above. But what I'd like to do is set one of the betas and
solve for the other
>
>> the seond in terms of the first. I also have a particular mean
>> structure I want to impose.

>
> hmm... for *interesting* conditions that come to my mind, I
> think of the means as being irrelevant.


Sorry. I don't follow this.
>
>

>> Can anyone point me to books or articles
>> outlining how to do this?
>>

>
>
> Maybe someone else isn't confused here, but for me, I
> can't tell what you want to do.


I hope I've clarified things enough for someone to provide some
references. Thanks in advance.





Point your RSS reader here for a feed of the latest messages in this topic.

[Privacy Policy] [Terms of Use]

© Drexel University 1994-2014. All Rights Reserved.
The Math Forum is a research and educational enterprise of the Drexel University School of Education.