Skip to content
Prev 201589 / 398506 Next

R: Re: R: Re: chol( neg.def.matrix ) WAS: Re: Choleski and Choleski with pivoting of matrix fails

Dear Peter,
thank you very much for your answer.

My problem is that I need to calculate the following quantity:

solve(chol(A)%*%Y)

Y is a 3*3 diagonal matrix and A is a 3*3 matrix. Unfortunately one 
eigenvalue of A is negative. I can anyway take the square root of A but when I 
multiply it by Y, the imaginary part of the square root of A is dropped, and I 
do not get the right answer.

I tried to exploit the diagonal structure of Y by using 2*2 matrices for A 
and Y. In this way the problem mentioned above disappears (since all 
eigenvalues of A are positive) and when I perform the calculation above I get 
approximately the right answer. The approximation is quite good. However it is 
an approximation.

Any suggestion?
Thank you very much!
Simon
with pivoting of matrix fails
it
eigen