Methods covered:
- Principle Component Analysis (PCA)
- Multidimensional Scaling (MDS)
- parametric (MDS)
- non-parametric (NMDS)
- Cluster Dendrogram
- Analysis of Similarity (ANOSIM)
The example dataset is dune from the vegan package.
- Dissimilarity matrix
Multidimensional scaling is a form of dimension reduction.
This table can be used as a rule of thumb.
| Stress | Meaning |
|---|---|
| 0.2 | poor fit |
| 0.1–0.2 | interpret cautiously |
| <0.1 | good ordination |
| <0.05 | excellent |
The analysis of similarity (ANOSIM) is a non-parametric test that uses a rank-based approach to test whether groups of samples are different.
- Technique explanations
- Add cluster analysis
- Add ANOSIM
- Expanded figure customisation
- Add table presentation for RMarkdown
- Improve biplot in PCA script