
Last chance! 50% off unlimited learning
Sale ends in
Tests for several groups of additive lognormally distributed compositions.
acompNormalLocation.test(x, g=NULL, var.equal=FALSE, paired=FALSE,
R=ifelse(var.equal,999,0))
a dataset of compositions (acomp) or a list of such
a factor grouping the data, not used if x is a list already.
Alternatively, g
can be a second compositional data set.
a boolean telling wether the variance of the groups should be considered equal
true if a paired test should be performed
number of replicates that should be used to compute p-values. 0 means comparing the likelihood statistic with the correponding asymptotic chisq-distribution.
A classical "htest"
object
The name of the dataset as specified
a name for the test used
an empty string
a dataset of p-value distributions under the Null-Hypothesis got from nonparametric bootstrap
The p.value computed for this test
Up to now the tests cannot handle missings.
The tests are based on likelihood ratio statistics.
Aitchison, J. (1986) The Statistical Analysis of Compositional Data Monographs on Statistics and Applied Probability. Chapman & Hall Ltd., London (UK). 416p.
# NOT RUN {
x <- runif.acomp(100,4)
y <- runif.acomp(100,4)
acompNormalLocation.test(list(x,y))
# }
Run the code above in your browser using DataLab