0

This graph shows the distribution of weeds on 3 treatments ( indicated by color). By googling I found that the graph is bounded between 0 and 1, where 0 means the sites (Treatment) have the same composition (that is they share all the species), and 1 means the sites do not share any species.

But what does x and y axis represents? Why does it require two dimensions?

Thanks

Mr Good News
  • 121
  • 1
  • 3
  • 15

1 Answers1

1

Check out this link.

NMDS ordination randomly places your data in 2D space, but then iteratively improves their ordinational distancing (spacing) to more closely represent their distance in the dissimilarity index. The axis are essentially arbitrary, but display your data in a way which best represents their dissimilarity. Points on the graph that are closer together are more similar (less dissimilar).

David Foster
  • 447
  • 4
  • 16
  • So it is put up in 2D space to improve their spacing?Thank you very much sir. – Mr Good News Feb 23 '18 at 10:08
  • I wouldn't say to improve the spacing, but to represent the dissimilarity of your samples across all of your variables in a simple visual format. You may want to do some form of indicator analysis to identify species that associate with your treatments (i.e. those that cause the separation of data on the graph). If so, check out `labdsv::indval()` – David Foster Feb 23 '18 at 10:48
  • 2
    NMDS has a clear way of thinking about this; it tries to position the sites in *k* dimensions such that the *rank order* of distances between sites in the ordination space are as a close as possible to the *rank order* of the original dissimilarities between sites. With "close as possible" being measured via the stress statistic. – Gavin Simpson Feb 23 '18 at 20:34
  • Why do you use NMDS? – Jari Oksanen Feb 24 '18 at 18:22
  • @JariOksanen I want to see the distribution of my samples in three treatments. We did some management practice and want to see if it had any effect on the weed community. And my dataset has null values. – Mr Good News Feb 26 '18 at 04:34