Lévy metric implementation

Asked Nov 15 '12 at 08:26

Active Feb 13 '14 at 15:46

Viewed 193 times

I would like to implement (in R) the evaluation of the Lévy distance between two curves. Given two cumulative distributions, the formal definition is the following (see also Wikipedia):

Essentially one is trying to find the largest square that can be drawed between the two curves.
I am dealing with two cumulatives which are step functions with same domain, so the implementation should be easier than the general case.
Thanks in advance.

Is there any library that implements this metric, or something similar. If not, could you suggest some hints to start from scratch?

edited Feb 08 '17 at 14:38

Community

asked Nov 15 '12 at 08:26

Mitch76

1

Is there a question here? If so, what have you tried? – Roman Luštrik Nov 15 '12 at 09:07
What Roman said. For example, can you program `inf(x)` , or the CDFs themselves? – Carl Witthoft Nov 15 '12 at 12:25
I edited the question, in order to be more clear, thanks. – Mitch76 Nov 22 '12 at 08:12

Lévy metric implementation

0 Answers0