search

shubhamraj / geoviz

Automatically exported from code.google.com/p/geoviz
0 stars 0 forks source link

Quantiles classification has problems #3

Open GoogleCodeExporter opened 9 years ago

GoogleCodeExporter commented 9 years ago 
What steps will reproduce the problem?
1. Load a dataset with many identical values
2. Note how the classification does not keep observations with the same
value in the same class.

Original issue reported on code.google.com by frank.hardisty on 2 Dec 2007 at 9:10

GoogleCodeExporter commented 9 years ago

Original comment by frank.hardisty on 2 Dec 2007 at 9:15

GoogleCodeExporter commented 9 years ago

Original comment by frank.hardisty on 3 Dec 2007 at 2:08

GoogleCodeExporter commented 9 years ago 
Comments from Alan

> I actually wonder whether something is not working right -- beyond
> the problem of dealing with skewed data.   I tried a 7 class raw
> quantile map with this and the US ends up in the 2nd highest category
> on both -- are there really that many zeros?  And, to confirm an
> apparent problem, when I tried modified quantiles, it shifted the US
> to the 3rd highest TOT but the 2nd lowest HOS ... of 7.

Original comment by frank.hardisty on 31 Jan 2008 at 2:19

GoogleCodeExporter commented 9 years ago 
Need more unit tests to verify that any fixes actually fix the problem, and to 
ensure
that the fixes do not introduce other problems.

Original comment by frank.hardisty on 31 Jan 2008 at 2:20