joefaith / targeted-projection-pursuit

Automatically exported from code.google.com/p/targeted-projection-pursuit
0 stars 0 forks source link

Visualise multiple classifications #11

Open GoogleCodeExporter opened 8 years ago

GoogleCodeExporter commented 8 years ago
TPP uses a single retinal variable to represent each possible classification 
(color, shape, size, fill). But this method cannot generalise beyond four 
possible classifications. There is a lot of interest at the moment in 
multi-label classification problems.

There is a discussion of recent research in the area here, including a Weka 
toolkit for multi-label problems:
http://mlkd.csd.auth.gr/multilabel.html

And a proposal for how these labels could be visualised is attached

Original issue reported on code.google.com by joedotfa...@gmail.com on 23 Jun 2011 at 9:20

GoogleCodeExporter commented 8 years ago
One way to show the relationship between label sets would be to use the 
notation of Haase diagrams: see http://en.wikipedia.org/wiki/Hasse_diagram

Original comment by joedotfa...@gmail.com on 6 Jul 2011 at 8:25