I'm not sure where the expression $

e_{i,-i}=frac{e_{i}}{1-H_{ii}}$came from, for $H_{ii}$ the i'th diagonal entry of the *hat matrix*. Could anyone please guide me to prove that the LHS would eventually yield the RHS? Thanks in advance!

### Similar Posts:

- Solved – why the diagonal elements of the inverted correlation matrix is related to correlation with all other variables
- Solved – First two moments of a quadratic form in which the vector and matrix are random (though independent)
- Solved – Marginal distribution of the diagonal of an inverse Wishart distributed matrix
- Solved – How to express a correlation matrix in terms of a covariance matrix
- Solved – Variance of a linear combination of vectors