corkdist {ade4}R Documentation

Tests of randomization between distances applied to 'kdist' objetcs

Description

The mantelkdist and RVkdist functions apply to blocks of distance matrices the mantel.rtest and RV.rtest functions.

Usage

mantelkdist (kd, nrepet = 999)
RVkdist (kd, nrepet = 999)
## S3 method for class 'corkdist':
plot(x, whichinrow = NULL, whichincol = NULL, 
   gap = 4, nclass = 10, coeff = 1,...)

Arguments

kd a list of class kdist
nrepet the number of permutations
x an objet of class corkdist, coming from RVkdist or mantelkdist
whichinrow a vector of integers to select the graphs in rows (if NULL all the graphs are computed)
whichincol a vector of integers to select the graphs in columns (if NULL all the graphs are computed)
gap an integer to determinate the space between two graphs
nclass a number of intervals for the histogram
coeff an integer to fit the magnitude of the graph
... further arguments passed to or from other methods

Details

The corkdist class has some generic functions print, plot and summary. The plot shows bivariate scatterplots between semi-matrices of distances or histograms of simulated values with an error position.

Value

a list of class corkdist containing for each pair of distances an object of class randtest (permutation tests).

Author(s)

Daniel Chessel
Stéphane Dray dray@biomserv.univ-lyon1.fr

Examples

data(friday87)
fri.w <- ktab.data.frame(friday87$fau, friday87$fau.blo, 
    tabnames = friday87$tab.names)
fri.kc <- lapply(1:10, function(x) dist.binary(fri.w[[x]],10))
names(fri.kc) <-  substr(friday87$tab.names,1,4)
fri.kd <- kdist(fri.kc)
 fri.mantel = mantelkdist(kd = fri.kd, nrepet = 999)
 plot(fri.mantel,1:5,1:5)
 plot(fri.mantel,1:5,6:10)
 plot(fri.mantel,6:10,1:5)
 plot(fri.mantel,6:10,6:10)
s.corcircle (dudi.pca(as.data.frame(fri.kd), scan = FALSE)$co)
plot(RVkdist(fri.kd),1:5,1:5)

data(yanomama)
m1 <- mantelkdist(kdist(yanomama),999)
m1
summary(m1)
plot(m1)

Worked out examples


> library(ade4)
> ### Name: corkdist
> ### Title: Tests of randomization between distances applied to 'kdist'
> ###   objetcs
> ### Aliases: corkdist mantelkdist RVkdist print.corkdist summary.corkdist
> ###   plot.corkdist
> ### Keywords: nonparametric
> 
> ### ** Examples
> 
> data(friday87)
> fri.w <- ktab.data.frame(friday87$fau, friday87$fau.blo, 
+     tabnames = friday87$tab.names)
> fri.kc <- lapply(1:10, function(x) dist.binary(fri.w[[x]],10))
> names(fri.kc) <-  substr(friday87$tab.names,1,4)
> fri.kd <- kdist(fri.kc)
>  fri.mantel = mantelkdist(kd = fri.kd, nrepet = 999)
>  plot(fri.mantel,1:5,1:5)
>  plot(fri.mantel,1:5,6:10)
>  plot(fri.mantel,6:10,1:5)
>  plot(fri.mantel,6:10,6:10)

> s.corcircle (dudi.pca(as.data.frame(fri.kd), scan = FALSE)$co)

> plot(RVkdist(fri.kd),1:5,1:5)

> 
> data(yanomama)
> m1 <- mantelkdist(kdist(yanomama),999)
> m1
Mantel's tests for 'kdist' object
class: corkdist list 
Call: mantelkdist(kd = kdist(yanomama), nrepet = 999)

gen-geo 
Monte-Carlo test
Call: mantelkdist(kd = kdist(yanomama), nrepet = 999)

Observation: 0.5098684 

Based on 999 replicates
Simulated p-value: 0.002 
Alternative hypothesis: greater 

     Std.Obs  Expectation     Variance 
3.248151e+00 1.174819e-05 2.463904e-02 

ant-geo 
Monte-Carlo test
Call: mantelkdist(kd = kdist(yanomama), nrepet = 999)

Observation: 0.8428053 

Based on 999 replicates
Simulated p-value: 0.001 
Alternative hypothesis: greater 

     Std.Obs  Expectation     Variance 
 5.212112779 -0.003004595  0.026334065 

ant-gen 
Monte-Carlo test
Call: mantelkdist(kd = kdist(yanomama), nrepet = 999)

Observation: 0.2995506 

Based on 999 replicates
Simulated p-value: 0.046 
Alternative hypothesis: greater 

    Std.Obs Expectation    Variance 
1.768687094 0.001214258 0.028451829 
list of 3 'randtest' objects
> summary(m1)
Mantel's tests for 'kdist' object
Call: mantelkdist(kd = kdist(yanomama), nrepet = 999)
Simulated p-values:
        1     2 3
geo     -     - -
gen 0.002     - -
ant 0.001 0.046 -
> plot(m1)

> 
> 
> 
>
[Package ade4 Index]