
Run t.test on each indiv value of x against all its neighbours (=remaining values of same vector) in order to test if tis value is likely to belong to vector x. This represents a repeated leave-one-out testing. Mutiple choices for multiple testing correction are available.
tTestAllVal(
x,
alph = 0.05,
alternative = "two.sided",
p.adj = NULL,
silent = FALSE,
debug = FALSE,
callFrom = NULL
)
This function returns a numeric vector with p-values or FDR (depending on argument p.adj
)
matrix or data.frame
(numeric) threshold alpha (passed to t.test
)
(character) will be passed to t.test
as argument 'alternative', may be "two.sided",..
(character) multiple test correction : may be NULL (no correction), "BH","BY","holm","hochberg" or "bonferroni" (but not 'fdr' since this may be confounded with local false discovery rate), see p.adjust
(logical) suppress messages
(logical) additional messages for debugging
(character) allow easier tracking of messages produced
set.seed(2016); x1 <- rnorm(100)
allTests1 <- tTestAllVal(x1)
hist(allTests1,breaks="FD")
Run the code above in your browser using DataLab