0

Right above this link here it's mentioned that the matrix of regression coefficients, $(\sum_{XX})^{-1}\sum_{XY}$, corresponds to the coefficients of an OLS solution. But I can't really see it for myself.

ChuckP
  • 743
  • 7
  • 17
  • See https://stats.stackexchange.com/search?q=multiple+regression+normal+equation+-normalization+-logistic. – whuber Nov 30 '17 at 15:14
  • I don't see how those questions are similar to mine. Mine is about the covariance-variance matrix of the dependent and independent variables relation to the OLS solution. – ChuckP Nov 30 '17 at 15:39
  • Those questions are *identical* to yours, unless you mean something unusual by "$\Sigma_{XX}$" and "$\Sigma_{XY}$". Feel free to edit your post to clarify the question and distinguish it from the others. – whuber Nov 30 '17 at 15:54
  • By those I mean the covariance matrix of the X vector and the covariance matrix(vector asssuming Y is 1-dim) of the vector of X and Y. Looking through the posts I've found this that resembles mine: https://stats.stackexchange.com/questions/3261/using-covariance-estimator-to-perform-linear-regression. but it does not have an answer as to why it(covariance estimator?) works as a estimator for OLS. – ChuckP Nov 30 '17 at 16:41
  • I added a third duplicate that might make the resemblance more apparent. – whuber Nov 30 '17 at 17:35
  • 1
    Somehow I did not see that one.. Thanks. – ChuckP Nov 30 '17 at 17:46

0 Answers0