distance-constructors: Distance Function Constructors

In bbuchsbaum/rMVPA: Multivoxel Pattern Analysis in R

Distance Function Constructors

Description

These convenience functions build specific types of distance function objects via create_dist. Each returns an S3 object inheriting from c("<method>", "distfun").

Usage

cordist(labels = NULL, method = c("pearson", "spearman"))

mahadist(labels = NULL)

eucdist(labels = NULL)

robustmahadist(labels = NULL)

pcadist(
  labels = NULL,
  ncomp = 2,
  whiten = TRUE,
  threshfun = NULL,
  dist_method = c("euclidean", "manhattan", "cosine")
)

Arguments

labels	Optional vector of row labels (not directly used in distance calculation).
method	For cordist, the correlation method: "pearson" or "spearman".
ncomp	For pcadist, the number of components (or a function threshold).
whiten	For pcadist, whether to whiten principal components (logical).
threshfun	For pcadist, an optional function that determines how many PCs to retain based on pres$sdev^2.
dist_method	For pcadist, the base distance measure in PC space ("euclidean", "manhattan", or "cosine").

Details

- cordist(labels, method="pearson") → correlation-based distance. - mahadist(labels) → Mahalanobis distance. - eucdist(labels) → Euclidean distance. - robustmahadist(labels) → Mahalanobis distance using robust covariance. - pcadist(labels, ...) → distance in reduced PCA space.

Value

An S3 object with class c("<method>", "distfun").

See Also

create_dist for the underlying constructor.

Examples

dist_obj_1 <- cordist(method="pearson")
dist_obj_2 <- mahadist()
dist_obj_3 <- eucdist()
dist_obj_4 <- robustmahadist()
dist_obj_5 <- pcadist(ncomp=2, dist_method="cosine")

bbuchsbaum/rMVPA documentation built on June 10, 2025, 8:23 p.m.