| Author |
 Topic  |
|
|
leonardo.a027
Argentina
9 Posts |
 Posted - 02/14/2017 : 12:44:24 AM
|
Hi.
I just made some 2d kernel density estimations of some data. Now I want to put them in the same scale to compare them, the problem is that the boundaty of the colored map is fixed and when I change the scale, the boudary of the colormap ends abruptaly:
how can I do to expand the boundary of the kernel density? |
Edited by - leonardo.a027 on 02/14/2017 12:47:16 AM |
|
|
Hideo Fujii
USA
1582 Posts |
Posted - 02/14/2017 : 11:03:15 AM
|
Hi leonardo.a027,
When you make a 2D Kernel Density plot, Origin creates a matrix of the density. You can expand the
range of the matrix by extrapolation with the menu, "Analysis: Mathematics: 2D Interpolate/Extrapolate".
However, to prevent the unintended result of the extrapolation, expanding should be not too much.
(For the best result, you can choose the inter/extrapolation method in that tool.) For the purpose of
aligning the ranges of two plots, you might also consider to shrink (=interpolate) one axis, instead.
Hope this helps.
--Hideo Fujii
OriginLab |
 |
|
|
leonardo.a027
Argentina
9 Posts |
Posted - 02/14/2017 : 12:25:32 PM
|
Hi Hideo.
Thanks a lot for the help.
I tried what you suggested and the result is the next:
As you can see, there are some irregular parts in the top part of the plot. I assume they are due to the extrapolation matrix and the length that the matrix was extrapolated.
I wonder if there's a way to expand the matrix range while doing the 2D kernel estimations, that way those irregularities will be avoided.
Thanks again. |
|
|
|
Hideo Fujii
USA
1582 Posts |
Posted - 02/14/2017 : 1:47:33 PM
|
Hi leonardo.a027,
I will ask our developers if it is possible to specify the XY ranges at the time to create a 2D Kernel Density
plot in the future release. Currently it's not possible.
Anyway, if I were you, I may consider the following:
As you notice that the original Z's range starts at 0.012, and your figure seems well-behaved with monotonic
declining from a single peak, try the Bilinear inter/extrapolation method; then set the color scale's bottom level
to 0.012 so that it shows that the expanded area also goes below that level. Thus, you don't want to allow
the irregular fluctuation below that level.
Just a possibility.
--Hideo Fujii
OriginLab |
Edited by - Hideo Fujii on 02/14/2017 1:48:14 PM |
|
|
|
leonardo.a027
Argentina
9 Posts |
Posted - 02/14/2017 : 1:59:48 PM
|
Thanks a lot!
I tried what you suggested and I got:
This will help me do what I intended :)
And I believe it will be very helpfull to include in future versions the possibility to specify the XY range in the 2Dkernel density estimations.
Thanks a lot again :) |
|
|
|
Hideo Fujii
USA
1582 Posts |
|
| |
Topic |
|
2D kernel density boundary