On Tue, Mar 16, 2010 at 5:03 PM, Raul Miller <[email protected]> wrote:
> On Tue, Mar 16, 2010 at 2:38 PM, Brian Schott <[email protected]> wrote:
>> I am not sure what you are asking, but I have struggled with dyadic I.
>> in its development of histograms because its direct usage is "biased"
>> oppositely to the statistical histogram. At the following  link I
>> said, "I. collects frequency counts based on intervals which are open
>> on the left and closed on the right: (xi-1,xi]. Statistical intervals
>> reverse this pattern: [xi-1,xi). "
>>
>> http://www.jsoftware.com/jwiki/BrianSchott/Histogram
>
> Does this mean that you want some definition for F such that
>   2 3 5 F 0 1 2 3 4 5 6 7
> _1 _1 0 1 1 2 2 2
>
> If so, here is a definition for F:
>
> F=: I. - ] < I. { [,_:

I finally noticed your definition for Idotr on your
Histogram page:
   Idotr=: |....@[ (#...@[ - I.) ]

Note that this is equivalent to
   IdotR=: 1 + F

Note also that for the integer domain this is equivalent
to
  Ier=: I. >:

For a floating domain replace >: with an operation
which adds epsilon to the right argument.  But your
sample data on your histogram page was integer.

FYI,

-- 
Raul
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm

Reply via email to