Suppose I have a multiple linear regression model $$ Y=\beta_0+\beta_1X_1+\cdots+\beta_pX_p+\epsilon$$ How can I obtain the regression coefficients $\hat{\beta_i}$ by fitting just a series of simple linear regression models (additionally using sample means of the response/predictor variables if needed)?
Asked
Active
Viewed 43 times
0
-
Why would you? If this is homework, please use the self-study tag. Hint: you could search this site for "Frisch Waugh Lovell". – Christoph Hanck Feb 28 '19 at 11:02
-
Alternatively, look into Gram-Schmidt orthogonalization. – Glen_b Feb 28 '19 at 12:12