Pcapro

From Eigenvector Documentation Wiki

Contents 1 Purpose 2 Synopsis 3 Description 3.1 Inputs 3.2 Outputs 4 See Also

Purpose

Project new data onto an existing principal components model.

Synopsis

[scoresn,resn,tsqn] = pcapro(newdata,loads,ssq,reslm,tsqlm,plots)

[scoresn,resn,tsqn] = pcapro(newdata,pcamod,plots)

Description

This function applies a previously-determined PCA model to a set of new data newdata. The PCA model can be input in one of two possible forms: 1) as a list of input variables, or 2) as a single model structure variable that had been previously returned by analysis or pca.

For case 1), the scaling for newdata should be the same as for the original data used to build the model. For case 2), pcapro will scale newdata based on data contained in pcamod

Inputs

Case 1) model input as a list of variables:

• newdata = data to be applied to the existing PCA model, scaled the same as the original data used to build the model
• loads = the model loadings
• ssq = the model variance information
• reslm = the limit for Q residuals
• tsqlm = the limit for T²
• plots = optional variable, which suppresses plotting when set to 0 {default plots = 1}.

Case 2) model input as a single model structure:

• newdata = data to be applied to the existing PCA model, in the units of the original data
• pcamod = the structure variable that contains the PCA model
• plots = optional variable, which suppresses the plots when set to 0 {default plots = 1}.

Outputs

• scoressn = the new scores
• resn = new residuals
• tsqn = new T² values

See Also

datahat, analysis, explode, modlpred, pca, simca, tsqmtx