[Rd] least squares with solve: leading minor is not positive definite

Tamas K Papp tpapp at Princeton.EDU
Wed Nov 29 17:29:29 CET 2006


I am following the "Comparing Least Squares Calculations" vignette of
the Matrix package to get a least squares solution x for

MM %*% x == mu$utility

I get the following error message:

Error in solve(crossprod(MM), crossprod(MM, mu$utility)) : 
	the leading minor of order 1079 is not positive definite

I thought that the least squares was always defined...  What can I do
about this?


More information about the R-devel mailing list