Principal components and principal coordinates: a comparative study based on an implementation to numerical taxonomy
Arce, Osvaldo E. A.; Nora E. De Marco; María R. Santillán. 2009. “Principal components and principal coordinates: a comparative study based on an implementation to numerical taxonomy”. Lilloa 46 (1-2). The objective of the present paper is to compare ordinations obtained from principal components and principal coordinates using a mixed data matrix corresponding to the Argentinean taxa of Echinochloa (Poaceae) under different application conditions. The following coefficients were used in principal coordinates: Euclidean distance, Manhattan, Bray Curtis and Canberra dissimilarities, and Gower similarity coefficient. Unstandardized and range or standard deviation standardized data were used. Ordination comparisons were accomplished using several techniques. All the analyses were run on the package NTSys. Corrections for negative eigenvalues were applied when necessary by means of Lingoes and Cailliez methods. Using Shepard diagrams and matrix to matrix correlations was very useful in order to judge ordinations. Standardization was the most important element to obtain appropriate ordinations. Gower coefficient handled appropriately the variables mixed nature. No important distortions in reduced dimensionality spaces were obtained when negative eigenvalues were present.
| Main Authors: | , , |
|---|---|
| Format: | Digital revista |
| Language: | spa |
| Published: |
Fundación Miguel Lillo
2009
|
| Online Access: | http://www.lillo.org.ar/journals/index.php/lilloa/article/view/446 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|