R: Distance Functions for nearest neighbours

ab.dist

R Documentation

Distance Functions for nearest neighbours

Description

These functions compute Absolute (ab.dist), Euclidean (eu.dist) or Mahalanobis (mh.dist) distances between two points. The variant functions (*.matY), accomplish the same task, but between a point on the one hand, and every point specified as rows of a matrix on the other.

Usage

ab.dist(x, y)
eu.dist(x, y)
mh.dist(x, y, A)
ab.dist.matY(x, Y)
eu.dist.matY(x, Y)
mh.dist.matY(x, Y, A)

Arguments

`x`	The vector (point) from which distance is sought.
`y`	The vector (point) to which distance is sought.
`Y`	A set of points, specified as rows of a matrix, to which distances are sought.
`A`	The inverse matrix to use for the Mahalanobis distance.

Details

These functions are used internally to decide how the nearest neighbours shall be calculated; the user need not call any of these functions directly. Rather, the choice of distance is specified as a string ('euclidean' or 'absolute' or 'mahal').

Value

Either a single number for the distance, or a vector of distances, corresponding to each row of Y.

Author(s)

Mohit Dayal

Examples

x <- c(1,2)
y <- c(0,3)
mu <- c(1,3)
Sigma <- rbind(c(1,0.2),c(0.2,1))
Y <- MASS::mvrnorm(20, mu = mu, Sigma = Sigma)
ab.dist(x,y)
eu.dist(x,y)
mh.dist(x,y,Sigma)
ab.dist.matY(x,Y)
eu.dist.matY(x,Y)
mh.dist.matY(x,Y,Sigma)