Please enable JavaScript to view this site.

This guide is for an old version of Prism. Browse the latest version or update Prism

Navigation: STATISTICS WITH PRISM 9 > Principal Component Analysis

Understanding Principal Component Analysis

Scroll Prev Top Next More

Principal Component Analysis (PCA) is a multivariate technique that is used to reduce the dimension of a dataset while retaining as much information from the data as possible. The pages of this section will provide a bit of background information on some of the concepts and techniques that are used when performing PCA. Although Prism will perform all of the "heavy lifting" in terms of data processing and computation, understanding the basic principles of the concepts involved can be extremely helpful when interpreting PCA results. If you want to skip the theory and get straight to analyzing data, this section of the guide will give you information on each of the options available for the analysis, and this section will help you understand the results that PCA generates.

Concepts behind PCA

Dimensionality reduction

Feature selection and feature extraction

Projecting data into lower dimensions

The PCA Process

Preparing data for analysis

Principal Components are defined using variance

Prinipcal components are orthogonal

Eigenvalues and eigenvectors

Selection of components

Classic methods for selecting PCs

Parallel Analysis

A complete example

Tabular results

Eigenvalues

Loadings (and eigenvectors)

PC Scores

Loadings Plot

PC Score Plots

Biplot

Eigenvalues (Scree) Plot

Proportion of variance plot

© 1995-2019 GraphPad Software, LLC. All rights reserved.