6

How can I fit reduced-rank regression with continuous response in R?

I found the package VGAM but it only fits for discrete distributions...

amoeba
  • 93,463
  • 28
  • 275
  • 317
Daniel Falbel
  • 570
  • 1
  • 3
  • 15
  • I do not know R, but reduced-rank regression has an explicit solution via standard regression and SVD, so it should not be difficult to implement manually. – amoeba Nov 27 '14 at 00:07
  • 3
    I'd be surprised is VGAM didn't do this; it has plenty of continuous distribution family functions (though note I haven't looked in detail at the RRR function in VGAM recently). You can also do something that is known as reduced rank regression with the **vegan** package. We call this Redundancy Analysis (RDA) but it also goes by the name reduced rank regression. And as @amoeba says, RDA can be computed by doing `fit – Gavin Simpson Nov 27 '14 at 14:29
  • 1
    @amoeba we may be talking about slightly different methods - RDA gets called a lot of things. We implement it in `rda()` via QR decomposition and SVD for efficiency, but *that* method gets the same result as the R code I showed in the comment earlier. Which makes me think what we do, which has been called reduced rank regression, is not the reduced rank regression the OP is looking for :-) – Gavin Simpson Nov 27 '14 at 18:03
  • Thank you all, I'll try to use this. I'm still trying to understand the rank reduced model – Daniel Falbel Nov 28 '14 at 11:02

2 Answers2

2

A set of S functions for least-squares reduced-rank can be found in the StatLib archive. See the file rrr.s and this paper:

Splus function for reduced-rank regression and softly shrunk reduced-rank regression. Submitted by Magne Aldrin (magne.aldrin@nr.no). [19/Apr/99][8/Mar/00] (14k)

Nir
  • 171
  • 4
rcs
  • 542
  • 1
  • 6
  • 22
2

There are now R packages for reduced-rank regression: rrpack, rrr.

Nir
  • 171
  • 4