You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I do not really like feature matrices since then we would have to know what others can do.
I think we should focus on what we can do, and do that in a more interesting looking form, i.e., a hand-written list of features. In fact, this should not be a list but more a general overview of what is possible -- nice to read, and well structured. We can then grow this, and also link against the notebooks or so.
I will try do start something before the release, at least for the new features.
Ok, if people like that, we should do it.
I guess I am just concerned about the usefulness of plain lists of things. What does model-selection (x) mean?
Also seems a lot of work to maintain such lists. But if its easy to extend, it is a good thing to do.
I will also try to write a short paragraph on each feature/machine and explain some contexts and problems for what it can be used. Like in the "Implemented method" in doxygen, but much less list-like/technical and more high-level.
Scikit learn website is a great example of useful feature lists
End of 2009 we had this
http://www.shogun-toolbox.org/jmlr10/
and we have a python script to generate this from a google docs .csv almost automagically. We need to display the tons of new features on the website.
The text was updated successfully, but these errors were encountered: