2

I have been working through the derivation of the formula used to calculate the Variance Inflation Factor associated with a model. I am hoping to start with the Least Squares equation as defined in matrix form and find a proof that derives this, linked here: derivation of VIF

I understand correlation is equal to cov/ $\hat{\sigma}^2$ and $VIF_{j}$ is the jth predictor is jth diagonal entry of inverse of correlation matrix. But how is this related to $VIF_{j}=\frac{Var({\hat{\beta_j}})}{\sigma^2}$ ?

I'd like to understand how to get from the matrix form of the least squares equation where $\hat{\beta}$ = $(X^TX)^{-1}(X^TY)$ to $VIF_j=\frac{Var({\hat{\beta_j}})}{\sigma^2}$???

Honestly, this is my main question and everything else is secondary. THANK YOU to anyone who can help me here!:)

From scanning multiple stack overflow questions, it seems that I can deduce that $X^TY$ = variance? And that $(X^TX)^{-1}$ is the correlation matrix or the standardized covariance matrix? (see highlighted portions #1 and #2). However, I'm not able to prove this or find information online that would lend insight on this issue.

I would really love if someone explained or gave insight on either question! Thank you so much!

Multivac
  • 2,784
  • 2
  • 8
  • 26
Erin
  • 81
  • 4
  • 1
    The image you added is not clear. If possible try adding Latex equations to your question for better understandability and reach... – Pluviophile Dec 06 '21 at 17:05
  • 1
    Can you elaborate on what is not clear? Are you having trouble reading it? If I transformed my work to Latex, would you be able to speak to the two questions I ask? Namely, whether the inverse of (X^T*X) is the correlation matrix and the relationship between the variance and X^T*Y? – Erin Dec 06 '21 at 18:09

0 Answers0