Orthogonal Distance Regression in 3d

Question

Does anyone know how to find a best fitting line for a point pattern in 3d?
I'm interested in the so called ODR (Orthogonal Distance Regression) Line, that minimizes the sum of the euclidian distances of all points to the line.

Is there a function in r for this problem or - if not - can you tell me how I can implement this myself?

Thanks in advance!

Answer 1

Try http://tolstoy.newcastle.edu.au/R/help/05/07/8039.html

Apparently there's a package, prcomp, which may meet your needs.