|
PureBytes Links
Trading Reference Links
|
> -----Original Message-----
> From: Chris Emery [mailto:cemery1@xxxxxxxxxxxxx]
> Sent: Saturday, August 31, 2002 5:15 AM
> To: omega-list@xxxxxxxxxx
> Subject: Correlation of three data series
>
>
> Dear List,
> Is there a formula for correlating 3 data series?
> I guess one could take the correlation of datas 1&2, 1&3 and 2&3, weight
> them according to which combinations do not contain the dependant variable
> but I am unsure if this would be the correct way to go about it. Any
> thoughts would be greatly appreciated.
You might want to take a look at multiple linear regression. The idea is
that you have a model where the dependent variable (Y) is a function of two
(or more) dependent variables:
Y = m1*X1 + m2*X2 ... +mn*Xn + b
For example, you might try expressing tomorrow's change in the S&P based on
today's change in the S&P (X1), today's change in interest rates (X2), and
today's change in VIX (likely an inverse relationship).
Excel can perform a multivariable regression using its analysis toolpack.
All this said, the problem is that even if there might be a relationship
between for example, the variables mentioned above and tomorrow's S&P, the
relationship might not be linear, and the relationship might not be stable.
|