dudi.hillsmith {ade4}R Documentation

Ordination of Tables mixing quantitative variables and factors

Description

performs a multivariate analysis with mixed quantitative variables and factors.

Usage

dudi.hillsmith(df, row.w = rep(1, nrow(df))/nrow(df), 
   scannf = TRUE, nf = 2)

Arguments

df a data frame with mixed type variables (quantitative and factor)
row.w a vector of row weights, by default uniform row weights are used
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

Details

If df contains only quantitative variables, this is equivalent to a normed PCA.
If df contains only factors, this is equivalent to a MCA.

This analysis is the Hill and Smith method and is very similar to dudi.mix function. The differences are that dudi.hillsmith allow to use various row weights, while dudi.mix deals with ordered variables.
The principal components of this analysis are centered and normed vectors maximizing the sum of :
squared correlation coefficients with quantitative variables
correlation ratios with factors

Value

Returns a list of class mix and dudi (see dudi) containing also
index a factor giving the type of each variable : f = factor, q = quantitative
assign a factor indicating the initial variable for each column of the transformed table
cr a data frame giving for each variable and each score:
the squared correlation coefficients if it is a quantitative variable
the correlation ratios if it is a factor

Author(s)

Stephane Dray dray@biomserv.univ-lyon1.fr
Anne B Dufour dufour@biomserv.univ-lyon1.fr

References

Hill, M. O., and A. J. E. Smith. 1976. Principal component analysis of taxonomic data with multi-state discrete characters. Taxon, 25, 249-255.

See Also

dudi.mix

Examples

   data(dunedata)
   attributes(dunedata$envir$use)$class <- "factor"   # use dudi.mix for ordered data
   dd1 <- dudi.hillsmith(dunedata$envir, scann = FALSE)
   scatter.dudi(dd1, clab.r = 1, clab.c = 1.5)
  

Worked out examples


> library(ade4)
> ### Name: dudi.hillsmith
> ### Title: Ordination of Tables mixing quantitative variables and factors
> ### Aliases: dudi.hillsmith
> ### Keywords: multivariate
> 
> ### ** Examples
> 
>    data(dunedata)
>    attributes(dunedata$envir$use)$class <- "factor"   # use dudi.mix for ordered data
>    dd1 <- dudi.hillsmith(dunedata$envir, scann = FALSE)
>    scatter.dudi(dd1, clab.r = 1, clab.c = 1.5)
> 
> 
> 
> 

[Package ade4 Index]