| Title: | Generalized Fisher's Combination Tests Under Dependence |
|---|---|
| Description: | Accurate and computationally efficient p-value calculation methods for a general family of Fisher type statistics (GFisher). The GFisher covers Fisher's combination, Good's statistic, Lancaster's statistic, weighted Z-score combination, etc. It allows a flexible weighting scheme, as well as an omnibus procedure that automatically adapts proper weights and degrees of freedom to a given data. The new p-value calculation methods are based on novel ideas of moment-ratio matching and joint-distribution approximation. The technical details can be found in Hong Zhang and Zheyang Wu (2020) <arXiv:2003.01286>. |
| Authors: | Hong Zhang and Zheyang Wu |
| Maintainer: | Hong Zhang <[email protected]> |
| License: | GPL-2 |
| Version: | 0.2.0 |
| Built: | 2025-01-04 02:53:22 UTC |
| Source: | https://github.com/cran/GFisher |
Survival function of the generalized Fisher's p-value combination statistic.
p.GFisher(q, df, w, M, p.type = "two", method = "HYB", nsim = NULL)p.GFisher(q, df, w, M, p.type = "two", method = "HYB", nsim = NULL)
q |
- observed GFisher statistic. |
df |
- vector of degrees of freedom for inverse chi-square transformation for each p-value. If all df's are equal, it can be defined by the constant. |
w |
- vector of weights. |
M |
- correlation matrix of the input statistics. |
p.type |
- "two" = two-sided p-values, "one" = one-sided p-values. |
method |
- "MR" = simulation-assisted moment ratio matching, "HYB" = moment ratio matching by quadratic approximation, "GB" = Brown's method with calculated variance. See details in the reference. |
nsim |
- number of simulation used in the "MR" method, default = 5e4. |
p-value of the observed GFisher statistic.
Hong Zhang and Zheyang Wu. "Accurate p-Value Calculation for Generalized Fisher's Combination Tests Under Dependence", <arXiv:2003.01286>.
set.seed(123) n = 10 M = matrix(0.3, n, n) + diag(0.7, n, n) zscore = matrix(rnorm(n),nrow=1)%*%chol(M) pval = 2*(1-pnorm(abs(zscore))) gf1 = stat.GFisher(pval, df=2, w=1) gf2 = stat.GFisher(pval, df=1:n, w=1:n) p.GFisher(gf1, df=2, w=1, M=M, method="HYB") p.GFisher(gf1, df=2, w=1, M=M, method="MR", nsim=5e4) p.GFisher(gf2, df=1:n, w=1:n, M=M, method="HYB") p.GFisher(gf2, df=1:n, w=1:n, M=M, method="MR", nsim=5e4)set.seed(123) n = 10 M = matrix(0.3, n, n) + diag(0.7, n, n) zscore = matrix(rnorm(n),nrow=1)%*%chol(M) pval = 2*(1-pnorm(abs(zscore))) gf1 = stat.GFisher(pval, df=2, w=1) gf2 = stat.GFisher(pval, df=1:n, w=1:n) p.GFisher(gf1, df=2, w=1, M=M, method="HYB") p.GFisher(gf1, df=2, w=1, M=M, method="MR", nsim=5e4) p.GFisher(gf2, df=1:n, w=1:n, M=M, method="HYB") p.GFisher(gf2, df=1:n, w=1:n, M=M, method="MR", nsim=5e4)
P-value of the omnibus generalized Fisher's p-value combination test.
p.oGFisher(p, DF, W, M, p.type = "two", method = "HYB", combine = "cct", nsim = NULL)p.oGFisher(p, DF, W, M, p.type = "two", method = "HYB", combine = "cct", nsim = NULL)
p |
- vector of input p-values. |
DF |
- matrix of degrees of freedom for inverse chi-square transformation for each p-value. Each row represents a GFisher test. |
W |
- matrix of weights. Each row represents a GFisher test. |
M |
- correlation matrix of the input statistics. |
p.type |
- "two" = two-sided p-values, "one" = one-sided p-values. |
method |
- "MR" = simulation-assisted moment ratio matching, "HYB" = moment ratio matching by quadratic approximation, "GB" = Brown's method with calculated variance. See details in the reference. |
combine |
- "cct" = oGFisher using the Cauchy combination method, "mvn" = oGFisher using multivariate normal distribution. |
nsim |
- number of simulation used in the "MR" method, default = 5e4. |
1. p-value of the oGFisher test. 2. individual p-value of each GFisher test.
Hong Zhang and Zheyang Wu. "Accurate p-Value Calculation for Generalized Fisher's Combination Tests Under Dependence", <arXiv:2003.01286>.
set.seed(123) n = 10 M = matrix(0.3, n, n) + diag(0.7, n, n) zscore = matrix(rnorm(n),nrow=1)%*%chol(M) pval = 2*(1-pnorm(abs(zscore))) DF = rbind(rep(1,n),rep(2,n)) W = rbind(rep(1,n), 1:10) p.oGFisher(pval, DF, W, M, p.type="two", method="HYB", combine="cct")set.seed(123) n = 10 M = matrix(0.3, n, n) + diag(0.7, n, n) zscore = matrix(rnorm(n),nrow=1)%*%chol(M) pval = 2*(1-pnorm(abs(zscore))) DF = rbind(rep(1,n),rep(2,n)) W = rbind(rep(1,n), 1:10) p.oGFisher(pval, DF, W, M, p.type="two", method="HYB", combine="cct")
Generalized Fisher's p-value combination statistic.
stat.GFisher(p, df = 2, w = 1)stat.GFisher(p, df = 2, w = 1)
p |
- vector of input p-values. |
df |
- vector of degrees of freedom for inverse chi-square transformation for each p-value. If all df's are equal, it can be defined by the constant. |
w |
- vector of weights. |
GFisher statistic sum_i w_i*qchisq(1 - p_i, df_i).
Hong Zhang and Zheyang Wu. "Accurate p-Value Calculation for Generalized Fisher's Combination Tests Under Dependence", <arXiv:2003.01286>.
n = 10 pval = runif(n) stat.GFisher(pval, df=2, w=1) stat.GFisher(pval, df=rep(2,n), w=rep(1,n)) stat.GFisher(pval, df=1:n, w=1:n)n = 10 pval = runif(n) stat.GFisher(pval, df=2, w=1) stat.GFisher(pval, df=rep(2,n), w=rep(1,n)) stat.GFisher(pval, df=1:n, w=1:n)