powerCal {riskPredictClustData} | R Documentation |
Calculate the power for testing δ=0
Description
Calculate the power for testing δ=0
.
Usage
powerCal(
nSubj,
mu1,
triangle,
rho,
rho11,
rho22,
rho12,
p11,
p10,
p01,
alpha = 0.05)
Arguments
nSubj |
integer. number of subjects to be generated. Assume each subject has two observations.
|
mu1 |
μ1=H(Y)−H(Yc) is the difference between probit transformation
H(Y) and probit-shift alternative H(Yc) ,
where Y is the prediction score of a randomly selected progressing
subunit, and Yc is the counterfactual random variable
obtained if each subunit that had progressed actually had not progressed.
|
triangle |
the difference of the expected value the the extended Mann-Whitney U statistics
between two prediction rules, i.e., △=ηc(1)−ηc(2)
|
rho |
ρ=corr(H(Zij),H(Zkℓ)) , where H=Φ−1 is the probit transformation.
|
rho11 |
ρ11=corr(Hij(1),Hiℓ(1)) , where H=Φ−1 is the probit transformation.
|
rho22 |
ρ22=corr(Hij(2),Hiℓ(2)) , where H=Φ−1 is the probit transformation.
|
rho12 |
ρ12=corr(Hij(1),Hiℓ(2)) , where H=Φ−1 is the probit transformation.
|
p11 |
p11=Pr(δi1=1&δi2=1) , where δij=1 if the j -th subunit of the
i -th cluster has progressed.
|
p10 |
p10=Pr(δi1=1&δi2=0) , where δij=1 if the j -th subunit of the
i -th cluster has progressed.
|
p01 |
p01=Pr(δi1=0&δi2=1) , where δij=1 if the j -th subunit of the
i -th cluster has progressed.
|
alpha |
type I error rate
|
Value
the power
Author(s)
Bernard Rosner <stbar@channing.harvard.edu>,
Weiliang Qiu <Weiliang.Qiu@gmail.com>,
Meiling Ting Lee <MLTLEE@umd.edu>
References
Rosner B, Qiu W, and Lee MLT.
Assessing Discrimination of Risk Prediction Rules in a Clustered Data Setting.
Lifetime Data Anal. 2013 Apr; 19(2): 242-256.
Examples
set.seed(1234567)
mu1 = 0.8
power = powerCal(nSubj = 30, mu1 = mu1,
triangle = 0.05, rho = 0.93, rho11 = 0.59, rho22 = 0.56, rho12 = 0.52,
p11 = 0.115, p10 = 0.142, p01 = 0.130, alpha = 0.05)
print(power)
[Package
riskPredictClustData version 0.2.6
Index]