ordinate {ordr}

Fit an ordination model to a data object

Description

This is a convenience function to fit an ordination model to a data object, wrap the result as a tbl_ord, and annotate this output with metadata from the model and possibly from the data.

Usage

ordinate(x, model, ...)

## Default S3 method:
ordinate(x, model, ...)

## S3 method for class 'array'
ordinate(x, model, ...)

## S3 method for class 'table'
ordinate(x, model, ...)

## S3 method for class 'data.frame'
ordinate(x, model, cols, augment, ...)

## S3 method for class 'dist'
ordinate(x, model, ...)

Arguments

x

A data object to be passed to the model, such as an array, table, data.frame, or stats::dist.

model

An ordination function whose output is coercible to class 'tbl_ord', or a symbol or character string (handled by match.fun()). Alternatively, a formula ~ fun(., ...) where fun is such a function and other arguments are explicit, which will be evaluated with x in place of ..

...

Additional arguments passed to model.

cols

<tidy-select> If x is a data frame, columns to pass to model. If missing, all columns are used.

augment

<tidy-select> If x is a data frame, columns to augment to the row data of the ordination. If missing, all columns not included in cols will be augmented.

Details

The default method fits the specified model to the provided data object, wraps the result as a tbl_ord, and augments this output with any intrinsic metadata from the model via augment_ord().

The default method is used for most classes, though this may change in future. The data.frame method allows the user to specify what columns to include in the model and what columns with which to annotate the output.

Value

An augmented tbl_ord.

Examples

# LRA of arrest data
ordinate(USArrests, cols = c(Murder, Rape, Assault), lra)

# CMDS of inter-city distance data
ordinate(UScitiesD, cmdscale_ord, k = 3L)

# PCA of iris data
ordinate(iris, princomp, cols = -Species, augment = c(Sepal.Width, Species))
ordinate(iris, cols = 1:4, ~ prcomp(., center = TRUE, scale. = TRUE))

# CA of hair & eye color data
haireye <- as.data.frame(rowSums(HairEyeColor, dims = 2L))
ordinate(haireye, MASS::corresp, cols = everything())

# FA of Swiss social data
ordinate(swiss, model = factanal, factors = 2L, scores = "Bartlett")

# LDA of iris data
ordinate(iris, ~ lda_ord(.[, 1:4], .[, 5], ret.x = TRUE))

# CCA of savings data
ordinate(
  LifeCycleSavings[, c("pop15", "pop75")],
  # second data set must be handled as an additional parameter to `model`
  y = LifeCycleSavings[, c("sr", "dpi", "ddpi")],
  model = cancor_ord, scores = TRUE
)