Polypls

Purpose

Calculate partial least squares regression models with polynomial inner relations.

Synopsis

[p,q,w,t,u,b,ssqdif] = polypls(x,y,lv,n)

Description

POLYPLS creates a partial least squares regression model with polynomial fit for the inner relation.

Outputs are p the x-block latent variable loadings, q the y-block variable loadings, w the x-block latent variable weights, t the x-block latent variable scores, u the y-block latent variable scores, b a matrix of polynomial coefficients for the inner relationship, and ssqdif a table of x- and y-block variance captured by the PLS model.

Use POLYPRED to make predictions with new data.

Inputs

• x = matrix of predictor variables (X-block),
• y = vector or matrix of the predicted variables (Y-block),
• lv = maximum number of latent variables to consider,
• n = order of polynomial to use for the inner-relation.

Outputs

• p = x-block loadings,
• q = y-block loadings,
• w = x-block weights,
• t = x-block scores,
• u = y-block scores,
• b = matrix of inner-relation coefficients,
• ssq = table of variance explained per component.

Options

options = a structure array with the following fields:

• plots: [ {'none'} | 'final' ] governs plotting of results, and
• order: positive integer for polynomial order {default = 1}.

See Also

lwrxy, pls, polypred