Can Theil-Sen be defined for multiple linear regression? If so, is there an implementation in R for it?

I simply want a formula a~b+c, but the package mblm fails with the (misleading) error message: stop("Only linear models are accepted"). Of course, this is a linear model: but they mean that Theil-Sen only works for two-dimensional linear models.

  • $\begingroup$ There's more than one way to generalize it. See here $\endgroup$ – Glen_b -Reinstate Monica Nov 3 '15 at 2:08
  • $\begingroup$ You may also like to consider the answer here $\endgroup$ – Glen_b -Reinstate Monica Nov 3 '15 at 2:15
  • $\begingroup$ I'm not sure this is a duplicate: neither of those include an R implementation $\endgroup$ – rhombidodecahedron Nov 3 '15 at 13:19
  • $\begingroup$ That part of your post is most likely off-topic instead of a duplicate (see the help center in the section under "Programming"). However, the answer at the second link definitely does mention an R package with an implementation for FastS (which isn't really Theil-Sen). I'm not aware of implementations in R for the methods at the first link. $\endgroup$ – Glen_b -Reinstate Monica Nov 3 '15 at 14:14