2

In statistics, what is an intuitive way of saying what an $n$-cube is? how do support of distributions come to be defined by an $n$-cube? are $n$-cubes useful in any applications beyond being a descriptor of dimensionality?

Richard Hardy
  • 54,375
  • 10
  • 95
  • 219
develarist
  • 3,009
  • 8
  • 31

1 Answers1

3

A unit 1-cube is the interval $[0,1]$ (or $(0,1)$ if you want an open rather than a closed interval) - i.e. a 1-cube is a line segment.

A unit 2-cube is the square region $[0,1] \times [0,1]$ (ditto, mutatis mutandis)

A unit 3-cube is the cubical volume $[0,1]^3$ (etc)

Hopefully the connection to the support of a distribution is clear, a unit $k$-cube would be the support for any continuous $k$-vector where each component takes values on the unit interval. For example, since you seem to be looking at copulas a lot lately, presumably you'll recognize that the support of a $k$-variate copula is the unit $k$-cube.

Glen_b
  • 257,508
  • 32
  • 553
  • 939