1

I have little knowledge about algebra, and need to rescale a precision(not variance-covariance) matrix. Suppose I have two variables, X1, X2, and their precision matrix is [5, .7, .7, .2]. Because the "variance" of the first variable(X1) should be 1, I thought that I could have such variance by multiplying the precision matrix by .2, but I wanted to confirm with you if this approach is correct.

Also, I am wondering if I need to rescale covariances as well and if so, how to? I don't think I need to take care of the variance of the other variable (X2, that is, .2).

kjetil b halvorsen
  • 63,378
  • 26
  • 142
  • 467
Shinzi Katoh
  • 21
  • 1
  • 2
    A consideration of the units in which the precision components are expressed will be an excellent guide towards a correct solution. – whuber Oct 20 '17 at 23:29

0 Answers0