corregp (version 0.1.4)

summary.corregp: Summarizing Correspondence Regression

Description

Method to produce a summary of a correspondence regression.

Usage

"summary"(object, parm = NULL, contrib = NULL, nf = NULL, add_ci = FALSE, cl = 0.95, nq = TRUE, ...)

Arguments

object

The outout of a call to corregp (i.e. an object of class "corregp").

parm

The parameter for which to compute the contributions contrib. Can be either "y" for the Y contributions, "x" for the X contributions, "both" which can be abbreviated to "b", or a vector of term names in X. Defaults to "b".

contrib

The type of contributions to be computed: either from points to axes (absolute contributions) or from axes to points (squared correlations). The specification can be "pnts_to_axes" or "axes_to_pnts", "pts2axs" or "axs2pts", "p_a" or "a_p", or any other reasonable abbreviation.

nf

The number of dimensions to be retained in the reduced space. Defaults to all dimenions (no reduction).

add_ci

Logical specifying whether to compute confidence intervals for the eigenvalues (and eigenvalues only). Defaults to FALSE.

cl

The confidence level for the confidence intervals. Defaults to 0.95.

nq

Logical specifying whether to use a normal quantile (i.e. apply qnorm) in the computation of the confidence intervals. Defaults to TRUE. If FALSE, then the confidence intervals are computed directly with the quantile function.

...

Further arguments passed to or from other methods.

Value

An object of class "summary.corregp", providing a summary of a correspondence regression, i.e. a list with components:

formula

The formula specified to the formula argument in the call to corregp.

data

The name of the data frame specified to the data argument in the call to corregp.

part

The name of the factor specified to thepart argument in the call to corregp.

chi_squared

The chi-squared value of the correspondence regression.

phi_squared

The phi-squared value of the correspondence regression, i.e. the chi-squared value divided by N.

N

The total number of observations.

eigen

Depending on add_ci: if FALSE, a matrix of the actual eigenvalues, their percentages and cumulative percentages; if TRUE, a list of the actual eigenvalues, their percentages and cumulative percentages together with the lower and upper confidence limits for each.

y

If parm is "y" or "b". A list of components p_a for the absolute contributions and//or a_p for the squared correlations, depending on contrib.

x

If parm is "y", "b" or any of the term names in X. A list of components p_a for the absolute contributions and/or a_p for the squared correlations, depending in contrib.

See Also

corregp, print.summary.corregp.

Examples

data(HairEye)
haireye.crg <- corregp(Eye ~ Hair * Sex, data = HairEye, b = 3000)
summary(haireye.crg, add_ci = TRUE)
summary(haireye.crg, parm = "y", contrib = "pts_axs", nf = 2)