Use an existing PCA solution (either a PCA object from CreatePCATable or a loading table) to compute principal component scores on a new dataset.
Usage
ProjectPCA(
Data,
VarsToReduce = NULL,
PCAInput,
InputType = c("PCAObj", "LoadingTable"),
center = TRUE,
scale = TRUE
)Arguments
- Data
Data frame on which to project PCA scores.
- VarsToReduce
Optional character vector of variable names to use. If NULL, uses all variables that appear in both Data and the PCA solution.
- PCAInput
Either:
the full object returned by CreatePCATable (when InputType = "PCAObj"), or
a loading table like CreatePCATable()$LoadingTable (when InputType = "LoadingTable").
- InputType
One of "PCAObj" or "LoadingTable".
- center
Logical; only used when InputType is "LoadingTable". For "PCAObj", the centering choice is taken from PCAInput$ScaleParams$center and this argument is ignored (with a warning if it conflicts).
- scale
Logical; only used when InputType is "LoadingTable". For "PCAObj", the scaling choice is taken from PCAInput$ScaleParams$scale and this argument is ignored (with a warning if it conflicts).
Value
A list with:
- Scores
Data frame of projected PCA scores.
- CombinedData
Original Data with scores appended as new columns.
- LoadingsUsed
Matrix of loadings used for projection.
- PCAObj
PCA object used (if InputType is "PCAObj"), otherwise NULL.
- VarsUsed
Variables used from Data for projection.
- Center
Logical flag indicating whether centering was applied for projection.
- Scale
Logical flag indicating whether scaling was applied for projection.