Learn R Programming

fastmatrix (version 0.3-8)

minkowski: Computes the p-norm of a vector

Description

Computes a p-norm of vector x using BLAS. The norm can be the one (p = 1) norm, Euclidean (p = 2) norm, the infinity (p = Inf) norm. For other values p >= 1 the underlying Fortran code is based on ideas of BLAS Level 1.

Usage

minkowski(x, p = 2)

Arguments

x

a numeric vector.

p

a number, specifying the type of norm desired. Possible values include real number greater or equal to 1, or Inf, Default value is p = 2.

Value

The vector p-norm, a non-negative number.

Details

Method of minkowski calls BLAS functions dasum (p = 1), dnrm2 (p = 2), idamax (p = Inf). For other values, a Fortran subroutine using unrolled cycles is called.

Examples

Run this code
# NOT RUN {
# a tiny example
x <- rnorm(1000)
minkowski(x, p = 1)
minkowski(x, p = 1.5)
minkowski(x, p = 2)
minkowski(x, p = Inf)

x <- x / minkowski(x)
minkowski(x, p = 2) # equal to 1
# }

Run the code above in your browser using DataLab