1

Suppose I have a blackbox function that solves simple linear regression. Can I use this function to solve "multiple" linear regression? The blackbox computes the slope and intercept in a simple linear regression model.

Nick Cox
  • 48,377
  • 8
  • 110
  • 156
user3910307
  • 11
  • 1
  • Does the function output residuals? At any rate you can calculate them from the coefficients & original data - see http://stats.stackexchange.com/a/17410/17230. – Scortchi - Reinstate Monica Feb 12 '16 at 09:38
  • Yes it does (as you said it's obvious). Thanks I will read it. – user3910307 Feb 12 '16 at 09:47
  • You're welcome. "Partialling out" is the key term. – Scortchi - Reinstate Monica Feb 12 '16 at 09:49
  • 1
    This text will help as well I think. Yes? http://www.isid.ac.in/~abhiroop/Site/TEACHING_files/Lecture%206.pdf – user3910307 Feb 12 '16 at 09:54
  • 2
    I have answered one interpretation of this question--how to estimate the coefficients--in detail at http://stats.stackexchange.com/questions/46185/question-on-how-to-normalize-regression-coefficient/46508#46508 and http://stats.stackexchange.com/questions/17336/how-exactly-does-one-control-for-other-variables/113207#113207. But if "solve" also includes the *extremely important* aspect of deriving the covariance matrix of the coefficient estimates, then additional work needs to be done. Is that your meaning? If so, then exactly what information does your function output? – whuber Feb 12 '16 at 13:29

0 Answers0