suprow {ade4} | R Documentation |
performs projections of supplementary rows.
## S3 method for class 'coa': suprow(x, Xsup, ...) ## Default S3 method: suprow(x, Xsup, ...) ## S3 method for class 'pca': suprow(x, Xsup, ...)
x |
an object of class dudi |
Xsup |
an array with the supplementary rows (Xsup and x$tab have the same column number) |
... |
further arguments passed to or from other methods |
If suprow.default
is used, the column vectors of Xsup are projected without prior modifications onto the principal components of dudi with the scalar product associated to the row weightings of dudi.
returns a data frame containing the coordinates of the supplementary projections
Daniel Chessel
Anne B Dufour dufour@biomserv.univ-lyon1.fr
Gower, J. C. (1967) Multivariate analysis and multivariate geometry. The statistician, 17, 13–28.
data(euro123) par(mfrow = c(2,2)) w <- euro123[[2]] dudi1 <- dudi.pca(w, scal = FALSE, scan = FALSE) s.arrow(dudi1$c1, sub = "Classical", possub = "bottomright", csub = 2.5) s.label(suprow(dudi1,w), add.plot = TRUE, clab = 0.75) s.arrow(dudi1$c1, sub = "Without centring", possub = "bottomright", csub = 2.5) s.label(suprow.default(dudi1,w), clab = 0.75, add.plot = TRUE) triangle.plot(w, clab = 0.75, label = row.names(w), scal = FALSE) triangle.plot(w, clab = 0.75, label = row.names(w), scal = TRUE) data(rpjdl) rpjdl.coa <- dudi.coa(rpjdl$fau, scann = FALSE, nf = 4) rpjdl.coa$li[1:3,] suprow(rpjdl.coa,rpjdl$fau[1:3,])$lisup # the same data(deug) deug.dudi <- dudi.pca(df = deug$tab, center = deug$cent, scale = FALSE, scannf = FALSE) suprow(deug.dudi, deug$tab[1:3,])$lisup # the supplementary individuals are centered deug.dudi$li[1:3,] # the same