weight_binary: Weight for the Binary effect sizes

In OPWeight: Optimal p-value weighting with independent information

Description

Compute weight from the probability of the rank given the effect size for the binary effect size

Usage

weight_binary(alpha, et, m, m1, tail = 1L, delInterval = 0.001, ranksProb)

Arguments

alpha	Numeric, significance level of the hypothesis test
et	Numeric, mean effect size of the test statistics
m	Integer, totoal number of hypothesis test
m1	Integer, number of true alternative hypothesis
tail	Integer (1 or 2), right-tailed or two-tailed hypothesis test. default is right-tailed test.
delInterval	Numeric, interval between the delta values of a sequence. Note that, delta is a LaGrange multiplier, necessary to normalize the weight
ranksProb	Numeric vector of the ranks probability of the tests given the effect size

Details

If one wants to test

H_0: epsilon_i=0 vs. H_a: epsilon_i = epsilon,

then et and ey should be median or any discrete value of the test and filter effect sizes, respectively. This is called hypothesis testing for the Binary effect sizes. m1 can be estimated using qvalue from a bioconductor package qvalue.

Value

weight Numeric vector of normalized weight of the tests for the binary case

Author(s)

Mohamad S. Hasan, [e-mail]

See Also

prob_rank_givenEffect weight_continuous qvalue

Examples

# compute the probabilities of the ranks of a test being rank 1 to 100 if the
# targeted test effect is 2 and the overall mean filter effect is 1.
ranks <- 1:100
prob2 <- sapply(ranks, prob_rank_givenEffect, et = 2, ey = 1, nrep = 10000,
                              m0 = 50, m1 = 50)
# plot the prooabbility
plot(ranks, prob2)

# compute weight for the binary case
weight_bin <- weight_binary(alpha = .05, et = 1, m = 100, m1 = 50, tail=1,
                             delInterval = .0001, ranksProb = prob2)

# plot the weight
plot(ranks, weight_bin)

OPWeight documentation built on Nov. 8, 2020, 11:06 p.m.