| Title: | Multicovariance and Multicorrelation for p-Variables |
|---|---|
| Description: | Implements the multicorrelation coefficient for p-variables as described in Cankaya (2023). The package provides a numerically stable implementation using logarithmic transformations and a log-sum-exp approach to reduce numerical overflow and underflow when calculations involve a large number of variables. |
| Authors: | Mehmet Niyazi Cankaya [aut, cre] |
| Maintainer: | Mehmet Niyazi Cankaya <[email protected]> |
| License: | GPL-3 |
| Version: | 0.1.0 |
| Built: | 2026-09-16 13:08:16 UTC |
| Source: | https://github.com/cran/multiCorr |
Computes the multicorrelation coefficient for p variables.
multicorrelation(x, r = 1)multicorrelation(x, r = 1)
x |
A numeric matrix or data frame. |
r |
A positive numeric power parameter. Default is 1. |
Each variable is centered by its arithmetic mean and scaled by its mean absolute deviation from the mean. Logarithmic transformations and a log-sum-exp approach are used to improve numerical stability.
A numeric multicorrelation coefficient.
Cankaya, M. N. (2023). Multicovariance and Multicorrelation for p-variables. In Mathematical Methods for Engineering Applications (ICMASE 2022), pp. 273-284. Cham: Springer International Publishing.
set.seed(123) X <- matrix(rnorm(1000), nrow = 100, ncol = 10) multicorrelation(X) multicorrelation(X, r = 2)set.seed(123) X <- matrix(rnorm(1000), nrow = 100, ncol = 10) multicorrelation(X) multicorrelation(X, r = 2)