Cameron Bieganek

I also just noticed that there is an undocumented MetaDiGraphs. :)

I agree. This would be a nice feature for 1.6, since that will be the new Long-Term Support version.

> any updates on this? @logankilpatrick I think you tagged the wrong person.

> making a SVM probabilisitic is expensive and has issues Ah, right, that's true if you want to return a calibrated probability. I didn't do much research before creating this...

I kind of like the approach of overloading `transform` for this. (If that's possible given the current method table.) I think it has a couple of advantages: - It doesn't...

Thanks! That's a pretty good solution. However, I think it would be worth adding a method to the MLJ API that does this out of the box. I think something...

That sounds good! Although the `drop_level` case is a little tricky, since you might want to be able to control the dropped level on a per column basis. [Scikit-learn](https://scikit-learn.org/stable/modules/generated/sklearn.preprocessing.OneHotEncoder.html) allows...

> `insert!(df, row)` Shouldn't that be the following? ```julia insert!(df, i, row) ``` Otherwise it would be the same as `push!`.

Hmm, well I think my suggested API for the `replace` data frame methods is intuitive at least. And maybe I haven't been following closely enough, but parts of the current...

A nice side benefit would be that `replace!(::DataFrame, ...)` would probably return the modified data frame, rather than the modified columns. I currently have this in one of my functions:...