zikw: Modified Kruskal Wallis test (ZIKW) for zero-inflated data

In chvlyl/ZIR: Modified Wilcoxon Rank Test and Kruskal Wallis Test for Zero-Inflated Data

Description

Modified Kruskal Wallis test (ZIKW) for zero-inflated data

Usage

zikw(x, group, perm = FALSE, perm.n = 10000)

Arguments

x	a numeric vector of data values with groups defined by the group parameter, or a list of numeric data vectors with each vector in the list as one group.
group	a vector giving the groups for the corresponding elements of x. Ignored if x is a list.
perm	use permutations to calculate the pvalue
perm.n	number of permutations used for pvalue calculation

Value

modified Kruskal Wallis test statistic and pvalue

References

Wanjie Wang, Eric Z. Chen and Hongzhe Li (2015). Rank-based tests for compositional distributions with a clump of zeros. Submitted.

Examples

## x is a list
x <- list(group1 = c(rep(0,5),rlnorm(20, meanlog = 0, sdlog = 1)),
     group2=c(rep(0,10),rlnorm(20, meanlog = 1, sdlog = 1)),
     group3=c(rep(0,15),rlnorm(20, meanlog = 2, sdlog = 1)))
zikw(x, perm = FALSE)
## x is a vector
x <- c(c(rep(0,5),rlnorm(20, meanlog = 0, sdlog = 1)),
    c(rep(0,10),rlnorm(20, meanlog = 1, sdlog = 1)),
    c(rep(0,15),rlnorm(20, meanlog = 2, sdlog = 1)))
group <- c(rep('group1',25),rep('group2',30),rep('group3',35))
zikw(x, group, perm = FALSE)
## use permutations to calculate the pvalue
## Not run: zikw(x, group, perm = TRUE)

chvlyl/ZIR documentation built on Oct. 14, 2021, 12:50 a.m.