Talk:Levenberg–Marquardt algorithm: Difference between revisions

Revision as of 03:11, 21 February 2008

Revision as of 03:15, 21 February 2008

Line 48:

where <math>\Sigma_{JJ}</math> is the square diagonal matrix:

<math>\Sigma_{JJ} = \Sigma_J^T\Sigma_J</math>

with <math>\Sigma_J</math> the square diagonal matrix:

with

<math>\Sigma_J = [\mbox{diag}(J)]^{-\frac{1}{2}}</math>

andso that the scaled Jacobian, <math>\hat{J}</math>, is:

<math>\hat{J} = \Sigma_{JJ}J</math>