zmjones/rfss

"Exploratory Data Analysis using Random Forests"

rfss's Issues

proximity

the proximity matrix might be something that we should discuss and use in the example. if we mapped it over time or projected it using mds or something it could be interesting in discovering dependence (social/spatial/temporal) as well as outliers (observations not proximate to anything else). both of which would be useful for later modeling/theorizing.

3d plot

Zach -- I like the 3d plot (interaction.png), but I don't see where you made it. Is that made in R?

Typo - "obtaind"

In the section "Combining the Trees to a Forest," "obtaind" should of course be "obtained."

example

we need a new (preferably regression) example. i've looked through the last 2 years of the APSR and didn't see much. there is so much political theory in there! also they don't have a replication data policy like AJPS does.

i am not unopposed to keeping the state repression example of course, but one with interaction would nice of course.