# Table Pca.pdf

# Principal component analysis

## Overview

**Principal Component Analysis** (PCA) can be used to get a generalized distribution for a data set, and detect outliers. One potential use of this command would be running PCA on an Affymetrix QC Report, in order to look for patterns in the quality control data. It can be accessed by going to **Table | Principal Component Analysis**.

### Input Data Requirements

The **Principal Component Analysis** Table menu option can be performed on any Table object.

## Step 1: Select source table

The user will first be asked to choose the table to run PCA on:

## Step 2: Principal Component Analysis Options

The user is then presented with the "Principal Component Analysis" window:

The Rows on which to perform the analysis can be selected (**all**, **selected**, **visible**, or a pre-generated List**), as well as the columns on which to perform the analysis. **

### Options

**Component Number**: Specify the number of components to be generated (default = 2).**Group**: A Group can be selected from a column (this will automatically color the results using this column).**Scale variables**: Choose whether or not to scale the variables for PCA (recommended and on by default).**Output scores**: Choose whether to generate a Table with scores (checked by default).**Output loadings**: Choose whether to generate a loading plot and .PcaLoadings Table.**Output eigen values**: Choose whether to generate a Text object with the eigen values.**Calculate Hotelling T2**Choose whether to generate a T2 Hotelling ellipse using the Alpha level (0.05 by default).- The new table can be optionally named using the "Output table name" box.

Note: Setting the Component number to 3 will return a 3D scatter view for the scores table, while setting the Component number to 4 or more returns a pairwise scatter view for the scores table.

## Output Results

A new folder named **Pattern** will be generated and automatically opens two views under the PcaSores table, one for Table view, and one for PcaScores view.

## Example Usage

Giving the following starting table and PCA options:

Output results:
**2D PCA**

Setting the Component number to 3 automatically returns a 3DScatterView for the scores table.

**3DScatterView **