weight_byEffect_bin: Weight for the different effect sizes in a binary case

In mshasan/OPWpaper: Reproduce figures and simulations of OPWeight

Description

Compute the weights from the ranks probability given the test-effect sizes when the effect sizes are binary.

Usage

weight_byEffect_bin(i, alpha, null, m, tail = 1L, delInterval,
  covariateEffectVec, datByNull)

Arguments

i	Integer, i-th effect size of a vector of effects
alpha	Numeric, a significance level of the hypothesis tests
null	Numeric, proportion of true true null tests
m	Integer, total number of hypotheis tests
tail	Integer (1 or 2), right-tailed or two-tailed hypothesis test.
delInterval	Numeric, interval between the delta values of a sequence. Note that, delta is a Langrangian multiplier, necessary to normalize the weight
covariateEffectVec	A numeric vector of covariate-effect sizes
datByNull	A numeric matrix of the ranks pobabilities in which each column corresponds to an effect size

Details

This function compute the weights when the effect sizes are binary by applying the ranks probabilities of the different effect sizes. It applies the function function weight_binary from the R package OPWeight to comute the weights from a probability matirx.

Value

A numeics matrix of weights in which each column corresponds to an effect size

Author(s)

Mohamad S. Hasan, [e-mail]

See Also

ranksProb_byEffect weight_binary

Examples

# vector of effect sizes
covariateEffectVec <- c(1, 1.5, 2)

# compute probability matrix
ranksProb_byEffect <- sapply(1:length(covariateEffectVec), ranksProb_byEffect,
             null = .9, m = 100, covariateEffectVec = covariateEffectVec)

# compute weights
weightByEffect <- sapply(1:length(covariateEffectVec), weight_byEffect_bin,
                   alpha = .05, null = .9, m = 100, delInterval = .01,
                   covariateEffectVec = covariateEffectVec,
                   datByNull = ranksProb_byEffect)

mshasan/OPWpaper documentation built on March 3, 2021, 7:02 a.m.