"Chia C Chong" <[EMAIL PROTECTED]> wrote in message
news:<9rsn26$98h$[EMAIL PROTECTED]>...
> I am a beginner in the statistical analysis and hypothesis. I have 2
> variables (A and B) from an experiment that was observed for a certain
> period time. I need to form a statistical model that will model these two
> variables. As an initial step, I plot the histograms of A & B separately to
> see how the data were distributed. However, it seems that both A & B can't
> be easily described by a simple statistical distributions like Gaussian,
> uniform etc via visualisation. Hence, I proceeded to plot the
> Quantile-Quantile plot (Q-Q plot) and trying to the fit both A and B with
> some theoretical distributions (all distributions avaiable in Matlab!!).
> Again, none of the distributions seem can descibe then completely. Then I
> was trying to perform the Wilcoxon Rank Sum test.
WHY? What is it you're trying to find out?
> From the data, it seems
> that A & B might be correlated in som sense.
Can you be more specific? Are the variables observed together, and
related so that A(i) is correlated with B(i)?
In that case, use a procedure that deals with the pairing, rather
than tossing them at a technique that relies on their independence.
Are A and B serially correlated with themselves?
Are they cross-correlated at some lag?
Please be clearer.
> My question is, what can I purely rely on the Wilcoxon Rank Sum Test to find
> the parameters of the distributions that can describe A & B??
Even if A and B satisified all the assumptions for the test,
IT WILL NOT TELL YOU "the parameters of the distributions that
can describe A & B".
Again, what are you trying to achieve?
> How do perform
> test to see whether A & B are really correlated?? How if A or/and B are
> overlay of two or more distributions?? Can this test tell me?? What make
> thing more tricky is that clustering was also observed in both A & B.
>
> I really hope to get an idea how to start with the statistical analysis for
> this kind problem...#
Don't start with some ill-chosen procedure, and then try to commit
acts of mayhem on your data until it will fit in the box. Start with
the questions you're trying to find out about, along with what you
know about the situation and believe about the data.
So answer these questions
-"What do I know?"
(write a list... e.g. i) data are pairs observed over time, ii)... )
-"What do I believe or expect before I start?"
(e.g. i) data pairs will be correlated ii) likely serial correlation, iii)...)
-"What do I want to know?"
*THEN* worry about how to do it (what procedure to use).
The methodology should not be the starting point!
Glen
=================================================================
Instructions for joining and leaving this list and remarks about
the problem of INAPPROPRIATE MESSAGES are available at
http://jse.stat.ncsu.edu/
=================================================================
