Multivariate analysis of (co)variance on Coe objects

Share:

Description

Performs multivariate analysis of variance on PCA objects.

Usage

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
MANOVA(x, fac, test = "Hotelling", retain, drop, verbose)

## S3 method for class 'OpnCoe'
MANOVA(x, fac, test = "Hotelling", retain, drop,
  verbose = TRUE)

## S3 method for class 'OutCoe'
MANOVA(x, fac, test = "Hotelling", retain, drop,
  verbose = TRUE)

## S3 method for class 'PCA'
MANOVA(x, fac, test = "Hotelling", retain = 0.99, drop,
  verbose = TRUE)

Arguments

x

a Coe object

fac

a name of a colum in the $fac slot, or its id, or a formula

test

a test for manova ('Hotelling' by default)

retain

how many harmonics (or polynomials) to retain, for PCA the highest number of PC axis to retain, or the proportion of the variance to capture.

drop

how many harmonics (or polynomials) to drop

verbose

logical whether to print messages

Details

Performs a MANOVA/MANCOVA on PC scores. Just a wrapper around manova. See examples for multifactorial manova and summary.manova for more details and examples.

Value

a list of matrices of (x,y) coordinates.

Note

Needs a review and should be considered as experimental.

See Also

Other multivariate: CLUST, KMEANS, LDA, MANOVA_PW, PCA

Examples

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
# MANOVA
data(bot)
bot.p <- PCA(efourier(bot, 12))
MANOVA(bot.p, 'type')

data(olea)
op <- PCA(npoly(olea, 5))
MANOVA(op, 'domes')

 m <- manova(op$x[, 1:5] ~  op$fac$domes * op$fac$var)
 summary(m)
 summary.aov(m)

 # MANCOVA example
 # we create a numeric variable, based on centroid size
 bot %<>% mutate(cs=coo_centsize(.))
 # same pipe
 bot %>% efourier %>% PCA %>% MANOVA("cs")

Want to suggest features or report bugs for rdrr.io? Use the GitHub issue tracker.