intercorr_fleish: Fleishman's Third-Order Polynomial Transformation...
In SimMultiCorrData: Simulation of Correlated Data with Multiple Variable Types

Description Usage Arguments Value References See Also

This function contains Fleishman's third-order polynomial transformation intermediate correlation equations (Headrick & Sawilowsky, 1999, doi: 10.1007/BF02294317). It is used in findintercorr and findintercorr2 to find the intermediate correlation for standard normal random variables which are used in the Fleishman polynomial transformation. It can be used to verify a set of constants and an intermediate correlation satisfy the equations for the desired post-transformation correlation. It works for two or three variables. Headrick & Sawilowsky recommended using the technique of Vale & Maurelli (1983, doi: 10.1007/BF02293687), in the case of more than 3 variables, in which the intermediate correlations are found pairwise and then eigen value decomposition is used on the correlation matrix. Note that there exist solutions that yield invalid power method pdfs (see power_norm_corr, pdf_check). This function would not ordinarily be called by the user.

1	intercorr_fleish(r, c, a)

r

either a scalar, in which case it represents pairwise intermediate correlation between standard normal variables, or a vector of 3 values, in which case:

r[1]*r[2] = ρ_{z1,z2},\ r[1]*r[3] = ρ_{z1,z3},\ r[2]*r[3] = ρ_{z2,z3}

c

a matrix with either 2 or 3 rows, each a vector of constants c0, c1, c2, c3, like that returned by find_constants

a

a matrix of target correlations among continuous variables; if nrow(a) = 1, it represents a pairwise correlation; if nrow(a) = 2 or 3, it represents a correlation matrix between two or three variables

a list of length 1 for pairwise correlations or length 3 for three variables; if the inputs satisfy the equations, returns 0 for all list elements

Please see references for findintercorr_cont.

fleish, power_norm_corr, pdf_check, find_constants

SimMultiCorrData documentation built on May 2, 2019, 9:50 a.m.