u_fisherZ: Calculate Fisher's Z Transformation for Correlations

In eclust: Environment Based Clustering for Interpretable Predictive Models in High Dimensional Data

Description

Calculate Fisher's Z transformation for correlations. This can be used as an alternative measure of similarity. Used in the s_generate_data function

Usage

u_fisherZ(n0, cor0, n1, cor1)

fisherTransform(n_1, r1, n_2, r2)

Arguments

n0	number of unexposed subjects
cor0	correlation matrix of unexposed covariate values. Should be dimension pxp
n1	number of exposed subjects
cor1	correlation matrix of exposed covariate values. Should be dimension pxp
n_1	number of unexposed subjects
r1	correlation for unexposed
n_2	number of exposed subjects
r2	correlation for exposed

Value

a pxp matrix of Fisher's Z transformation of correlations

Note

fisherTransform is called internally by u_fisherZ function

References

https://en.wikipedia.org/wiki/Fisher_transformation

Examples

data("simdata")

X = simdata[,c(-1,-2)]
fisherScore <- u_fisherZ(n0 = 100, cor0 = cor(X[1:50,]),
                         n1 = 100, cor1 = cor(X[51:100,]))

dim(fisherScore)

fisherScore[1:5,1:5]

eclust documentation built on May 1, 2019, 8:46 p.m.