Hi

I don't believe it is posible to get a good estimate of Y as you know
nothing about it. Knowing that X is correlated with Y doesn't help much
except if you know HOW they are correlated!

I'm relatively new in this field, so please correct me if I'm wrong!

Thomas


"Alan Miller" <[EMAIL PROTECTED]> wrote in message
q9ui9.25102$[EMAIL PROTECTED]">news:q9ui9.25102$[EMAIL PROTECTED]...
> "Helen" <[EMAIL PROTECTED]> wrote in message
> news:amdk8r$9j8$[EMAIL PROTECTED]...
> > Hi there
> >
> > I have two random vector X (X1,X2,...,Xn) and Y (Y1,Y2,...,Yn)
> > I know there are correlated, now I have already known pdf of X,
> > what kind of ways could I get the pdf of Y?
> > And to access this aim, what kind of other information( maybe
> > covariance matrix etc.) should I know?
> > I am now building a model for my research, so I wanna get some clue
> > about that.
> >
> > Thanks a lot!
> >
> >
>
> For fairly narrow bands of X-values, I would like to look at histograms of
> the corresponding Y-values.   You may find that, say, the means and std.
> deviations both increase roughly linearly with the mean of the X-values.
> You may find that the shape of the histogram may look roughly the same for
> each group, but just moved.
>
> The most comprehensive book I know on bivariate distributions is:
> Hutchinson, T.P. and Lai, C.D. (1991) `The Engineering Statistician's
Guide
> to Continuous Bivariate Distributions', published by Rumsby Scientific
> Publishing.
> ISBN 0 646 02413 2
> Chapter 15 `How to understand bivariate data' attempts to answer questions
> such as yours.
> The book is highly mathematical, and much of it is devoted to cataloging a
> large number of bivariate distributions.
>
> Cheers
>
> --
> Alan Miller
> http://users.bigpond.net.au/amiller
> Retired Statistician
>
>
>


.
.
=================================================================
Instructions for joining and leaving this list, remarks about the
problem of INAPPROPRIATE MESSAGES, and archives are available at:
.                  http://jse.stat.ncsu.edu/                    .
=================================================================

Reply via email to