mahalanobis_distance_cpp: (Generalized) Mahalanobis distance
In shapr: Prediction Explanation with Dependence-Aware Shapley Values

mahalanobis_distance_cpp

R Documentation

(Generalized) Mahalanobis distance

Description

Used to get the Euclidean distance as well by setting mcov = diag(m).

Usage

mahalanobis_distance_cpp(
  featureList,
  Xtrain_mat,
  Xexplain_mat,
  mcov,
  S_scale_dist
)

Arguments

`featureList`	List. Contains the vectors indicating all factor combinations that should be included in the computations. Assumes that the first one is empty.
`Xtrain_mat`	Matrix Training data in matrix form
`Xexplain_mat`	Matrix Explanation data in matrix form.
`mcov`	matrix The covariance matrix of X.
`S_scale_dist`	logical. Indicating whether the Mahalanobis distance should be scaled with the number of variables

Value

Array of three dimensions. Contains the squared distance for between all training and test observations for all feature combinations passed to the function.