PCA analysis under PLS,PCA module extracts Eigenvectors based on "Correlations" or "Covariances" matrix.
Article ID: KB0083789
Updated On:
| Products | Versions |
|---|---|
| Spotfire Statistica | 13.3, 13.2, 13.1,13.0 |
Description
PCA analysis under menu "Statistics | Mult/Exploratory | Principal Components & Classification" module would extract eigenvectors/eigenvalues from either "Correlations" or "Covariances" matrix.
Does the PCA analysis under PLS,PCA module extract Eigenvectors based on "Correlations" and "Covariances" matrix as well?
Does the PCA analysis under PLS,PCA module extract Eigenvectors based on "Correlations" and "Covariances" matrix as well?
Issue/Introduction
This article shows guidance on the PCA analysis which extracts Eigenvectors based on "Correlations" or "Covariances" matrix.
Environment
Windows 7, Windows Server 2012 R2
Resolution
By default, PCA from the PLS, PCA module will extract the eigenvectors from the Correlations matrix.
In order to extract eigenvectors from the Covariance matrix, user needs to check the option “Do not scale variables” which is found on the Advanced tab.
User may (or may not) need to adjust the fitting options of the NIPALS algorithm, that is, number of iterations and convergence criterion found on the NIPALS tab, in order to have the results matched to the ones generated from the PCA analysis under "Statistics | Mult/Exploratory | PCA" module.
In order to extract eigenvectors from the Covariance matrix, user needs to check the option “Do not scale variables” which is found on the Advanced tab.
User may (or may not) need to adjust the fitting options of the NIPALS algorithm, that is, number of iterations and convergence criterion found on the NIPALS tab, in order to have the results matched to the ones generated from the PCA analysis under "Statistics | Mult/Exploratory | PCA" module.
Feedback
Yes
No
Powered by
