rosa_results: Result functions for ROSA models

In multiblock: Multiblock Data Fusion in Statistics and Machine Learning

Result functions for ROSA models

Description

Standard result computation and extraction functions for ROSA (rosa).

Usage

## S3 method for class 'rosa'
predict(
  object,
  newdata,
  ncomp = 1:object$ncomp,
  comps,
  type = c("response", "scores"),
  na.action = na.pass,
  ...
)

## S3 method for class 'rosa'
coef(object, ncomp = object$ncomp, comps, intercept = FALSE, ...)

## S3 method for class 'rosa'
print(x, ...)

## S3 method for class 'rosa'
summary(
  object,
  what = c("all", "validation", "training"),
  digits = 4,
  print.gap = 2,
  ...
)

blockexpl(object, ncomp = object$ncomp, type = c("train", "CV"))

## S3 method for class 'rosaexpl'
print(x, digits = 3, compwise = FALSE, ...)

rosa.classify(object, classes, newdata, ncomp, LQ)

## S3 method for class 'rosa'
scores(object, ...)

## S3 method for class 'rosa'
loadings(object, ...)

Arguments

object	A rosa object.
newdata	Optional new data with the same types of predictor blocks as the ones used for fitting the object.
ncomp	An integer giving the number of components to apply (cummulative).
comps	An integer vector giving the exact components to apply (subset).
type	For blockexpl: Character indicating which type of explained variance to compute (default = "train", alternative = "CV").
na.action	Function determining what to do with missing values in newdata.
...	Additional arguments. Currently not implemented.
intercept	A logical indicating if coefficients for the intercept should be included (default = FALSE).
x	A rosa object.
what	A character indicating if summary should include all, validation or training.
digits	The number of digits used for printing.
print.gap	Gap between columns when printing.
compwise	Logical indicating if block-wise (default/FALSE) or component-wise (TRUE) explained variance should be printed.
classes	A character vector of class labels.
LQ	A character indicating if 'max' (maximum score value), 'lda' or 'qda' should be used when classifying.

Details

Usage of the functions are shown using generics in the examples below. Prediction, regression coefficients, object printing and summary are available through: predict.rosa, coef.rosa, print.rosa and summary.rosa. Explained variances are available (block-wise and global) through blockexpl and print.rosaexpl. Scores and loadings have their own extensions of scores() and loadings() throught scores.rosa and loadings.rosa. Finally, there is work in progress on classifcation support through rosa.classify.

When type is "response" (default), predicted response values are returned. If comps is missing (or is NULL), predictions for length(ncomp) models with ncomp[1] components, ncomp[2] components, etc., are returned. Otherwise, predictions for a single model with the exact components in comps are returned. (Note that in both cases, the intercept is always included in the predictions. It can be removed by subtracting the Ymeans component of the fitted model.)

If comps is missing (or is NULL), coef()[,,ncomp[i]] are the coefficients for models with ncomp[i] components, for i = 1, \ldots, length(ncomp). Also, if intercept = TRUE, the first dimension is nxvar + 1, with the intercept coefficients as the first row.

If comps is given, however, coef()[,,comps[i]] are the coefficients for a model with only the component comps[i], i.e., the contribution of the component comps[i] on the regression coefficients.

Value

Returns depend on method used, e.g. predict.rosa returns predicted responses or scores depending on inputs, coef.rosa returns regression coefficients, blockexpl returns an object of class rosaexpl containing block-wise and component-wise explained variance contained in a matrix with attributes.

References

Liland, K.H., Næs, T., and Indahl, U.G. (2016). ROSA - a fast extension of partial least squares regression for multiblock data analysis. Journal of Chemometrics, 30, 651–662, doi:10.1002/cem.2824.

See Also

Overviews of available methods, multiblock, and methods organised by main structure: basic, unsupervised, asca, supervised and complex. Common functions for computation and extraction of results and plotting are found in rosa_results and rosa_plots, respectively.

Examples

data(potato)
mod <- rosa(Sensory[,1] ~ ., data = potato, ncomp = 5, subset = 1:20)
testset <- potato[-(1:20),]; testset$Sensory <- testset$Sensory[,1,drop=FALSE]
predict(mod, testset, ncomp=5)
dim(coef(mod, ncomp=5)) # <variables x responses x components>
print(mod)
summary(mod)
blockexpl(mod)
print(blockexpl(mod), compwise=TRUE)

multiblock documentation built on Nov. 18, 2023, 5:06 p.m.