| histogram bin edges [message #82555] |
Sat, 22 December 2012 18:56  |
Josh Sixsmith
Messages: 13
Registered: December 2012
|
Junior Member |
|
|
Hi, I'm curious about whether IDL is inclusive or exclusive in regards to the final bin. The help documentation describes that the right hand side of a bin is exclusive, but what about the final bin?
For example, using the following sample data
a = [ 0.72244781, 0.20885457, 0.38053078, 0.89579923, 0.93703798,$
1. , 0.22754776, 0.11365818, 0.38424101, 0.1741128 ,$
0.63094614, 0.00123615, 0.06025917, 0.78652067, 0.1001857 ,$
0.80492211, 0.80564817, 0.83369342, 0.94378603, 0.75453023]
ha = histogram(a, nbins=10, binsize=0.1, min=0, reverse_indices=ria, locations=loca, omax=omaxa, omin=omina)
IDL> print, loca
0.00000 0.10000 0.20000 0.30000 0.40000 0.50000
0.60000 0.70000 0.80000 0.90000
IDL> print, ha
2 3 2 2 0 0
1 3 4 3
This result looks fine and i interpret the locations as
[0,0.1),[0.1,0.2),[0.2,0.3),[0.3,0.4),[0.4,0.5),[0.5,0.6),[0 .6,0.7),[0.7,0.8),
[0.8,0.9),[0.9,1.0]
indicating the last bin as inclusive as the value of 1.0 is included in the histogram.
However i feel that this might actually be a case of misrepresentation of floating point numbers, as mentioned in the example given in http://www.idlcoyote.com/math_tips/razoredge.html
If array 'a' is double:
ad = double(a)
had = histogram(ad, nbins=10, binsize=0.1d, min=0.0d, reverse_indices=riad, locations=locad, omin=ominad, omax=omaxad)
print, had
2 3 2 2 0 0
1 3 4 2
which suggests that the last bin is exclusive and that all bins are exclusive
[0,0.1),[0.1,0.2),[0.2,0.3),[0.3,0.4),[0.4,0.5),[0.5,0.6),[0 .6,0.7),[0.7,0.8),
[0.8,0.9),[0.9,1.0)
For integer data this also seems to be the case
IDL> print, locbHowever, for integer data this doesn't seem to be the case.
b = indgen(11)
hb = histogram(b, nbins=10, min=0, reverse_indices=rib, locations=locb, omax=omaxb, omin=ominb)
IDL> print, locb
0 1 2 3 4 5 6 7 8 9
IDL> print, hb
1 1 1 1 1 1
1 1 1 1
IDL> print, rib
11 12 13 14 15 16
17 18 19 20 21 0
1 2 3 4 5 6
7 8 9
The index '10' for the value 10 is not included
So if i use the max keyword, i would assume that this would be an upper limit of values to be included in the histogram. Does this then make the last bin inclusive, or will it create a bin that will not only include the max value but potentially include values slightly higher than the specified max. This might only occur if the binsize is set.
Any clarification would be greatly appreciated.
Cheers
Josh
|
|
|
|
| Re: histogram bin edges [message #82621 is a reply to message #82555] |
Wed, 02 January 2013 02:46  |
Josh Sixsmith
Messages: 13
Registered: December 2012
|
Junior Member |
|
|
On Sunday, 23 December 2012 13:56:24 UTC+11, Josh Sixsmith wrote:
> Hi, I'm curious about whether IDL is inclusive or exclusive in regards to the final bin. The help documentation describes that the right hand side of a bin is exclusive, but what about the final bin?
>
>
>
> For example, using the following sample data
>
>
>
> a = [ 0.72244781, 0.20885457, 0.38053078, 0.89579923, 0.93703798,$
>
> 1. , 0.22754776, 0.11365818, 0.38424101, 0.1741128 ,$
>
> 0.63094614, 0.00123615, 0.06025917, 0.78652067, 0.1001857 ,$
>
> 0.80492211, 0.80564817, 0.83369342, 0.94378603, 0.75453023]
>
>
>
> ha = histogram(a, nbins=10, binsize=0.1, min=0, reverse_indices=ria, locations=loca, omax=omaxa, omin=omina)
>
> IDL> print, loca
>
> 0.00000 0.10000 0.20000 0.30000 0.40000 0.50000
>
> 0.60000 0.70000 0.80000 0.90000
>
>
>
> IDL> print, ha
>
> 2 3 2 2 0 0
>
> 1 3 4 3
>
>
>
> This result looks fine and i interpret the locations as
>
>
>
> [0,0.1),[0.1,0.2),[0.2,0.3),[0.3,0.4),[0.4,0.5),[0.5,0.6),[0 .6,0.7),[0.7,0.8),
>
> [0.8,0.9),[0.9,1.0]
>
>
>
> indicating the last bin as inclusive as the value of 1.0 is included in the histogram.
>
>
>
> However i feel that this might actually be a case of misrepresentation of floating point numbers, as mentioned in the example given in http://www.idlcoyote.com/math_tips/razoredge.html
>
>
>
> If array 'a' is double:
>
> ad = double(a)
>
> had = histogram(ad, nbins=10, binsize=0.1d, min=0.0d, reverse_indices=riad, locations=locad, omin=ominad, omax=omaxad)
>
> print, had
>
> 2 3 2 2 0 0
>
> 1 3 4 2
>
>
>
> which suggests that the last bin is exclusive and that all bins are exclusive
>
>
>
> [0,0.1),[0.1,0.2),[0.2,0.3),[0.3,0.4),[0.4,0.5),[0.5,0.6),[0 .6,0.7),[0.7,0.8),
>
> [0.8,0.9),[0.9,1.0)
>
>
>
> For integer data this also seems to be the case
>
> IDL> print, locbHowever, for integer data this doesn't seem to be the case.
>
> b = indgen(11)
>
> hb = histogram(b, nbins=10, min=0, reverse_indices=rib, locations=locb, omax=omaxb, omin=ominb)
>
> IDL> print, locb
>
> 0 1 2 3 4 5 6 7 8 9
>
> IDL> print, hb
>
> 1 1 1 1 1 1
>
> 1 1 1 1
>
>
>
> IDL> print, rib
>
> 11 12 13 14 15 16
>
> 17 18 19 20 21 0
>
> 1 2 3 4 5 6
>
> 7 8 9
>
> The index '10' for the value 10 is not included
>
>
>
> So if i use the max keyword, i would assume that this would be an upper limit of values to be included in the histogram. Does this then make the last bin inclusive, or will it create a bin that will not only include the max value but potentially include values slightly higher than the specified max. This might only occur if the binsize is set.
>
>
>
> Any clarification would be greatly appreciated.
>
>
>
> Cheers
>
> Josh
Great, thanks for that.
I'll keep that in mind for future use.
Cheers
Josh
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
