MANOVA {Momocs}R Documentation

Multivariate analysis of (co)variance on Coe objects

Description

Performs multivariate analysis of variance on PCA objects.

Usage

MANOVA(x, fac, test = "Hotelling", retain, drop)

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

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

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

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

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. Silent message and progress bars (if any) with options("verbose"=FALSE).

See Also

Other multivariate: CLUST(), KMEANS(), KMEDOIDS(), LDA(), MANOVA_PW(), MDS(), MSHAPES(), NMDS(), PCA(), classification_metrics()

Examples

# MANOVA
bot.p <- PCA(efourier(bot, 12))
MANOVA(bot.p, 'type')

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")


[Package Momocs version 1.4.1 Index]