Pcrengine

Purpose

Principal components regression computational engine.

Synopsis

[reg,ssq,loads,scores,pcassq] = pcrengine(x,y,ncomp,options)

Description

PCRENGINE calculates the basic elements of a PCR model (see PCR), given a matching set of x-block and y-block data belonging to the same set of samples.

Inputs

• x = the x-block data, and
• y = the corresponding y-block data.

Optional Inputs

• ncomp = the number of components to to be calculated (positive integer scalar). If the number of components ncomp is not specified, the routine will return components up to the rank of the x-block.
• options, which is discussed below.

Outputs

• reg = the matrix of regression vectors
• ssq = the sum of squares captured ssq
• loads = x-block loadings
• scores = x-block scores
• pcassq = the PCA sum-of-squares table

Options

options = a structure array with the following fields:

• display: [ 'off' | {'on'} ], governs level of display to command window,

• sortorder: [ {'x'} | 'y' ], governs order of factors in outputs. 'x' is standard PCR sort order (ordered in terms of X block variance captured). 'y' is Correlation PCR sort order (ordered in terms of Y block variance captured).

The default options can be retreived using: options = pcrengine('options');.

See Also

analysis, pcr, pls