DiscriMiner (version 0.1-22)

desDA: Descriptive Discriminant Analysis

Description

Performs a Descriptive Discriminant Analysis (a.k.a. Factorial Discriminant Analysis from the french Analyse Factorielle Discriminante)

Usage

desDA(variables, group, covar = "within")

Arguments

variables

matrix or data frame with explanatory variables

group

vector or factor with group memberships

covar

character string indicating the covariance matrix to be used. Options are "within" and "total"

Value

• An object of class "desda", basically a list with the following elements
• power
  table with discriminant power of the explanatory variables
• values
  table of eigenvalues
• discrivar
  table of discriminant variables, i.e. the coefficients of the linear discriminant functions
• discor
  table of correlations between the variables and the discriminant axes
• scores
  table of discriminant scores for each observation

Details

When covar="within" the estimated pooled within-class covariance matrix is used in the calculations. When covar="total" the total covariance matrix is used in the calculations. The difference between covar="within" and covar="total" is in the obtained eigenvalues.

The estiamted pooled within-class covariance matrix is actually the within-class covariance matrix divided by the number of observations minus the number of classes (see getWithin)

References

Lebart L., Piron M., Morineau A. (2006) Statistique Exploratoire Multidimensionnelle. Dunod, Paris.

See Also

discPower

Examples

# load bordeaux wines dataset
  data(bordeaux)

  # descriptive discriminant analysis with within covariance matrix
  my_dda1 = desDA(bordeaux[,2:5], bordeaux$quality)
  my_dda1

  # descriptive discriminant analysis with total covariance matrix
  my_dda2 = desDA(bordeaux[,2:5], bordeaux$quality, covar="total")
  my_dda2
  
  # plot factor coordinates with ggplot
  library(ggplot2)
  bordeaux$f1 = my_dda1$scores[,1]
  bordeaux$f2 = my_dda1$scores[,2]
  ggplot(data=bordeaux, aes(x=f1, y=f2, colour=quality)) + 
  geom_hline(yintercept=0, colour="gray70") +
  geom_vline(xintercept=0, colour="gray70") +
  geom_text(aes(label=year), size=4) + 
  opts(title="Discriminant Map - Bordeaux Wines (years)")