Last chance! 50% off unlimited learning
Sale ends in
performs a multivariate analysis with mixed quantitative variables and factors.
dudi.hillsmith(df, row.w = rep(1, nrow(df))/nrow(df),
scannf = TRUE, nf = 2)
Returns a list of class mix
and dudi
(see dudi) containing also
a factor giving the type of each variable : f = factor, q = quantitative
a factor indicating the initial variable for each column of the transformed table
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
a data frame with mixed type variables (quantitative and factor)
a vector of row weights, by default uniform row weights are used
a logical value indicating whether the eigenvalues bar plot should be displayed
if scannf FALSE, an integer indicating the number of kept axes
Stéphane Dray stephane.dray@univ-lyon1.fr
Anne-Béatrice Dufour anne-beatrice.dufour@univ-lyon1.fr
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
Hill, M. O., and A. J. E. Smith. 1976. Principal component analysis of taxonomic data with multi-state discrete characters. Taxon, 25, 249-255.
dudi.mix
data(dunedata)
attributes(dunedata$envir$use)$class <- "factor" # use dudi.mix for ordered data
dd1 <- dudi.hillsmith(dunedata$envir, scann = FALSE)
if(adegraphicsLoaded()) {
g <- scatter(dd1, row.plab.cex = 1, col.plab.cex = 1.5)
} else {
scatter(dd1, clab.r = 1, clab.c = 1.5)
}
Run the code above in your browser using DataLab