Multiple Regression
Author: Christos
Creation Date: 3/10/2011 7:28 AM
profile picture

Christos

#1
I would like to ask if there is any way to perform multiple linear regression in WL. Are there any addins or indicators? Secondly is there any way to convert DataSeries to Bars?
profile picture

Eugene

#2
QUOTE:
Secondly is there any way to convert DataSeries to Bars?

Yes there is but to determine the optimal one first let me ask what is the purpose?
profile picture

Christos

#3
I am trying to use the Beta indicator. The problem is that I want to run the calculation on the ROC^2 and not on the ROC which the beta indicator does by default. What would you suggest? How about my first question?
profile picture

Eugene

#4
QUOTE:
How about my first question?

When I have something to comment you can expect it to happen. Not this time though -- nothing to say.
QUOTE:
I am trying to use the Beta indicator. The problem is that I want to run the calculation on the ROC^2 and not on the ROC which the beta indicator does by default.

If you mean the Beta indicator from the Community.Indicators library which is designed to accept a Bars object, then I'm not sure how would conversion of DataSeries to Bars help this at all. (Just a note: the Beta internally does ROC^2 on the market DataSeries.) The code is open so you could modify the indicator as you please after downloading the project source code from the Wiki. Let me know if further assistance with this modification is required.
profile picture

sedelstein

#5
I'm not sure the question was answered here. Is there a way to perform a multiple linear regression?

I'd like to fit a polynomial of the form y = a +b1(t) + b2(t^2) + ....

Alternatively, has anyone called a matrix algebra package from within WL?
profile picture

Eugene

#6
QUOTE:
Alternatively, has anyone called a matrix algebra package from within WL?


Hope this helps:

Use external library (ALGLIB) in my code
error while using external DLL library
This website uses cookies to improve your experience. We'll assume you're ok with that, but you can opt-out if you wish (Read more).