7 comments

[ 3.9 ms ] story [ 24.5 ms ] thread
"Rattle" by Togaware seems a very easy to use Data Mining GUI for R. It includes Decision Trees.

However, I've found tutorials (for Rattle) hard to come by. If anyone wants to add their favorite tutorials here, please do so... it'll turn this page into a nice resource.

Agree. I've done a lot of decision trees in R lately and the two packages I've seen with the "most" documentation are rattle and rpart. I've found rpart much easier to use and understand, and it has pretty involved documentation. [1]

[1]http://cran.r-project.org/web/packages/rpart/vignettes/longi...

Does rpart have a GUI module?
Yup, you can run printcp, plotcp, and fancyRpartPlot on it to make it look nice.
I remember this game written in the "logo" language that could guess the animal you were thinking of, by guiding the user through a decision tree; and it also allowed the user to grow the tree when it failed at guessing.
Tree based methods, such as Random Forests, provide some nice properties at the expense of interpretability. Trees by themselves tend to be fairly variable.
Thank you for your inputs. I will be providing with forests article on the same