Mutual Information — mi • infoxtr

Estimate the mutual information between target and interacting variables.

Usage

mi(
  data,
  target,
  interact,
  base = exp(1),
  type = c("cont", "disc"),
  k = 3,
  normalize = FALSE
)

data: Observation data.
target: Integer vector of column indices for the target variables.
interact: Integer vector of column indices for the interacting variables.
base: (optional) Logarithm base of the entropy. Defaults to exp(1) (nats). Use 2 for bits or 10 for dits.
type: (optional) Estimation method: "disc" for discrete entropy or "cont" for continuous entropy (KSG estimator).
k: (optional) Number of nearest neighbors used by the continuous estimator. Ignored when type = "disc".
normalize: (optional) Logical; if TRUE, return normalized mutual information.

A numerical value.

infoxtr::mi(matrix(1:100,ncol=2),1,2)
#> [1] 2.979205