0

In a simple linear regression: Assume that the leverage of a specific data point equals 1 (maximum leverage),

from my understanding, it also means that the variance of the estimated error (e) of that particular data point equals to 0.

does it also mean that the value of e must equal to zero (the prediction matches perfectly the true y value)? or not necessarily?

Xi'an
  • 105,342
Alex Il
  • 45
  • 3
    When the variance of a random variable is zero, that variable is almost surely constant. The expectation of each error is assumed to be zero. The conclusion is easy to draw! – whuber Feb 20 '23 at 15:26
  • The proof is contained here: https://stats.stackexchange.com/questions/212656/leverages-and-effect-of-leverage-points – kjetil b halvorsen Mar 02 '23 at 18:59

0 Answers0