More Masturbating with the Data

[From Bill Williams 9 Feburary 2004 1:52 AM CST]

[From Rick Marken (2004.02.08.2250)]

> Bill Powers (2004.02.08.0734 MST)

> I'm totally confused. What are you modeling?

My question would be where is the economic content?
Or, am I missing something?

Perhaps modeling is too strong a word.

What you are doing has nothing to do with modelling.

All I was trying to do was get a

better sense of the meaning of my lagged correlation graphs.>

There is no reason that approaching the problem in the
way you are can't lead to some insights. But, you seem
to be approaching the problem with what amounts to
a random attack that is completely uninformed by any
reference to economic content. I suggest you read the
Peach and Webb paper on randomly specified models.

I hope this clarifies things a bit.

It clarifies things enomously- except where the economy went-- that you
don't explain.

As you can see I did fix up my

equations because there were errors in the earlier version. Comments
and suggestions would be most welcome.

Ask Professor Bruun for her estimate of the chance that this approach will
be likely to generateuseful results.

Read the Peach and Webb paper. It could save a lot of
wasted effort. Also, the last years Swedish bank prize was given for work
on the statistical issues involved in
the treatment of correlated variables-- as most of economic are, which makes
data mining an almost completely futile approach. Peach and Webb provide a
very accessible explanation of the issues regarding data mining.

Bill Williams