I analyze my ecological data using Canonical Correspondence Analysis (CCA) and I am a little bit confused with the eigenvalue function in ecological interpretation, is there an eigenvalue function in ecological data interpretation, or it is just statistical?
The eigenvalues of a correlation (or covariance) matrix correspond to the entity of variability along the principal axes (maximal inertia direction of motions) of a system. Their ecological interpretation stems from the analysis of loadings, see:
Article The Application of Principal Component Analysis to Drug Disc...
Hello Adhi,
the practical use of eigenvalues is to rank the corresponding scores in decreasing order of their role in building the factorial space. It is often expressed as the percentage of the total variance.
An important (mathematically) phenomenom will have a large eigenvalue.
But it often appears that these mathematically important phenomena are not the more important for your research ! It is up to you to find the more interesting subspace for your work.
Nevertheless, a very small eigenvalue (say a few percent of the total variance) may be an artefact.
Kind regards,
Dominique
https://www.freecodecamp.org/news/an-overview-of-principal-component-analysis-6340e3bc4073/
http://practicalcryptography.com/miscellaneous/machine-learning/guide-principal-component-analysis-pca/
- Badges
- Science topic
- Similar topics
- Counseling Psychology
- Confusion