Hei again,
just looked on it and it seems to be fine :)
I only changed the limits by calculating them between the actual and the
next item. Otherwise the limits will be "on" one item and my
classification algorithm will classify it wrong.
Tomorrow I will check (hopefully) against the results given in a text
book (T.A. Slocum, 1999). If that fits i will try to make a new version
on the weekend.
stefan
Hisaji ONO schrieb:
Hello.
Your plug-in is very nice.
I've added Jenks Optimization Method in your
plug-in(which I ported original Jenks's Basic program
which were available from Jenks's co-worker. Original
code(Fortran?) are listed in his technoical paper cited in
Slocum's book).
For this, I've needed modified and added code of
following files.
-org.openjump.core.classify.Classifier1D.java
-- public static String JENKS_BREAKS = "Jenks
Optimization"; <-- added
-- public static double[]
classifyNaturalBreaks(double[] data, int numberClasses)
<-- added
-- public static List
getAvailableClassificationMethods()
<-- modified
-org.openjump.core.ui.plugin.ClassifyAttributesPlugIn.java
-- private FeatureDataset classifyAndCreatePlot
<-- modified
-org.openjump.core.resources.classifyplot_ja_JP.properties
<-- added # Japanese Language Propeties file
-org.openjump.core.resources.classifyplot*.properties
--
org.openjump.core.classify.Classifier1D.Jenks-Breaks=Jenks
Optimisation Method <-- added
I've attached jar file, "dataexploration.jar", please
replace, try and evaluate it. If you like it, please add
your distribution.
I have three questions about your plug ins.
1. When choosing 3 as number of classes for all
classification methods, this plugins returned wrong value
for last(max.) break's values.
2. Like ESDA in GeoDa, have you planed to be enable to
highlight points when selecting some features(points,
linestrings. polygons) in map pane dynamically.
3. Not only selecting each point, but also selecting
points falled within circles or polygons drawn by mouse?
4. Have you planed classification methods will be called
from color theming panel from style dialog like SIGLE's
"Merge" OpenJUMP?
Regards.
--- Stefan Steiniger <[EMAIL PROTECTED]> wrote:
(note: I also tried to implement Jenks' optimal
breaks, but I could not
find a sufficient description of the method - only a
not feasable "try
all possible breaks" description. That is why I
implemented k-means for
a pseudo-optimization/fine tuning of the derived
class breaks from the
other methods.)
------------------------------------------------------------------------
_______________________________________________
jump-users mailing list
[email protected]
http://lists.refractions.net/mailman/listinfo/jump-users
_______________________________________________
jump-users mailing list
[email protected]
http://lists.refractions.net/mailman/listinfo/jump-users
