User Tools

Site Tools


en:pca_examples

Differences

This shows you the differences between two versions of the page.

Link to this comparison view

Both sides previous revision Previous revision
Next revision
Previous revision
en:pca_examples [2019/02/27 14:52]
David Zelený
en:pca_examples [2019/02/27 16:13]
David Zelený [Example 3: Evaluation of importance of ordination axes in PCA]
Line 64: Line 64:
 ==== Example 2: PCA on environmental matrix ==== ==== Example 2: PCA on environmental matrix ====
  
-[[en:​data:​wetlands|Carpathian wetlands dataset (Hájek et al.)]] contains information about species composition of vascular plants and mosses, and also extensive information about the environment,​ mostly water chemistry. In the following example, we will explore the intercorrelated nature of environmental ​dataset.+[[en:​data:​wetlands|Carpathian wetlands dataset (Hájek et al.)]] contains information about species composition of vascular plants and mosses, and also extensive information about the environment,​ mostly water chemistry. In the following example, we will explore the intercorrelated nature of environmental ​variables. Note that we select PCA since we assume linear correlations between the variables; there is not reason to apply DCA on the data first to decide between linear and unimodal ordination methods (this decision step is reserved only for species composition data).
  
 <code rsplus> <code rsplus>
Line 402: Line 402:
 {{:​obrazky:​ordination_unc19b.png?​direct|}} {{:​obrazky:​ordination_unc19b.png?​direct|}}
  
-You can see that ordination diagram looks ok - we may start to happily interpret it. But broken stick comparison shows that there is actually no axis worth to interpret - their eigenvalues are smaller than those of the null broken stick solution (broken stick interpretation applies only for the first axes; it may seem that broken stick indicates that short axes like 6, 7 and 8 may be interpreted,​ but it would not make sense).+You can see that ordination diagram looks ok - we may start to happily interpret it. But broken stick comparison shows that there is actually no axis worth to interpret - their eigenvalues are smaller than those of the null broken stick solution (broken stick interpretation applies only for the first axes; it may seem that broken stick indicates that short axes like 6, 7 and 8 may be interpreted,​ but it would not make any sense to interpret them).
  
  
en/pca_examples.txt · Last modified: 2019/02/27 16:13 by David Zelený