SensoMineR (version 1.03)

afmult: Multiple Factor Analysis

Description

Performs a multiple factor analysis, using an object of class ktab.

Usage

afmult(X, option = c("lambda1", "inertia", "uniform", "internal"),
     coord = c(1,2), scannf = TRUE, nf = 3, cex = 0.8, 
     col = "steelblue4", font = 2, clabel = 0.8, scale.unit = TRUE)

Arguments

X

an object of class ktab

option

a string of characters for the weighting of arrays options: lambda1{weighting of group k by the inverse of the first eigenvalue of the k analysis} inertia{weighting of group k by the inverse of

coord

a length 2 vector specifying the components to plot

scannf

a logical value indicating whether the eigenvalues bar plot should be displayed

nf

if scannf FALSE, an integer indicating the number of kept axes

cex

cf. function par in the graphics package

col

cf. function par in the graphics package

font

cf. function par in the graphics package

clabel

cf. the ade4 package

scale.unit

a boolean, if TRUE (value set by default) then data are scaled to unit variance

Value

• Returns a list including:
• tab
  a data frame with the modified array
• rank
  a vector of ranks for the analyses
• eig
  a numeric vector with the all eigenvalues
• li
  a data frame with the coordinates of rows
• TL
  a data frame with the factors associated to the rows (indicators of table)
• co
  a data frame with the coordinates of columns
• TC
  a data frame with the factors associated to the columns (indicators of table)
• blo
  a vector indicating the number of variables for each table
• lisup
  a data frame with the projections of normalized scores of rows for each table
• link
  a data frame containing the projected inertia and the links between the arrays and the reference array

Details

The only difference between this function and the mfa programed by Daniel Chessel lies in the calculations of the coordinates of the partial individuals.

References

Escofier, B. and Pag�s, J. (1994) Multiple factor analysis (AFMULT package), Computational Statistics and Data Analysis, 18, 121--140.

Examples

data(sensopanels)
resafmult<-afmult(ktab.data.frame(sensopanels, blocks = rep(14,7)),
    scannf = FALSE)