Re: Modelling data with cyclical patterns?
- From: "David Jones" <dajxxxx@xxxxxxxxx>
- Date: Tue, 19 Jun 2007 18:01:06 +0100
cody wrote:
I've already talked to the locally available professors, they don't
know anything about nonlinear regression. I have a masters degree in
statistics from a major university (not local).
My job includes managing the database and making statistical reports.
The last guy who had this job was a high school graduate who didn't
know what a primary key is, or what a standard deviation is. He died
of cancer, so we should see him in the database soon. I'm quite an
improvement over that.
FYI, I found out about functions such as nls in R or proc NLIN in
SAS. So I guess you can go back to making your cheapshot comments,
I'll worry more about saving children.
If you are the original poster, then please understand that this was not meant to be a "cheapshot comment" but
serious advice. I had assumed that you had some statistical knowledge which is why I suggested you should try to
find "a more experienced statistician". You should realise that the problem is not necessarily a regression one.
There are people who will have worked with similar types of data for many years and who will know what sorts of
things to include in the models and what sorts of models are useful. You might be able to pick up some of this
from books or journals, but why start from scratch? If, as you said, "answers will help policy planners in a
life-or-death sort of way" then presumably your employers should be prepared to bring in advisors from outside
who already have the experience. If you plan to be in the same or similar job for some time, then their
experience can rub off on you and your employers can benefit from this in future. They of course would want to
see the data, or relevant parts of statistical analyses at least.
If you can't find a statistician with relevant experience locally, you might see if you can find an actuary
working on life-tables, possibly for an insurance company, but this might only be done at HQ-type places.
Possibly others here can suggest appropriate terminology for your type of problem that you could use for a
literature search.
David Jones
.
- Follow-Ups:
- References:
- Modelling data with cyclical patterns?
- From: hp
- Re: Modelling data with cyclical patterns?
- From: David Jones
- Re: Modelling data with cyclical patterns?
- From: cody
- Modelling data with cyclical patterns?
- Prev by Date: Re: Modelling data with cyclical patterns?
- Next by Date: When is Poisson Regression Appropriate?
- Previous by thread: Re: Modelling data with cyclical patterns?
- Next by thread: Re: Modelling data with cyclical patterns?
- Index(es):
Relevant Pages
|