Tyler Tomita

Agreed. I would assume this is a requirement for integration with the caret package, which is what I ultimately would like to see happen at some point. On Mon, Jun...

Why integrate with caret? It makes model tuning much easier. It provides a uniform interface for comparing many different algorithms. Most of the well-known algorithms, including XGBoost and Rborist have...

My feeling is that splitting randomly on an end point will really hurt generalization, particularly for smaller sample sizes. > On Jul 26, 2018, at 1:14 PM, joshua vogelstein wrote:...

Consider class 0 a 1d Gaussian centered at -1 and class 1 a Gaussian centered at +1. Suppose our sample consists of one data point from each class at each...

It should definitely be an option but further tests should be done before deciding whether it’s the default. > On Jul 26, 2018, at 1:14 PM, joshua vogelstein wrote: >...

The things that readily come to mind are lasso or RF variable importance. I'll have to think about this some more.

Now that I think about it, I think F-RC needs to be included. With ROFLMAO published, I anticipate a lot more awareness of and interest in F-RC.

I will add it to all pertinent figures

i will add that to the outline

feature importance is easy because the set has cardinality p. it's not clear to me what a feasible way to plot subspace importance is because the set of all subspaces...