Heiko Strathmann
Heiko Strathmann
Start with one of the algos maybe. KNN or so
> @karlnapf the only issue will be how does the user have access to `compute_svm_primal_objective` and `compute_svm_dual_objective`? Maybe SVM shouldn't be removed from swig, but classes like `MKL` should register...
ah! crap! :) TBH I don't think it is such a problem.... I would actually just not expose it for now. But these cases require a solution eventually ...
they are computed anyways when fitting the model. But how would you compute them in apply? Would you store them? Dont think that is any different from storing the labels
Just realised, the solution is an observer
@geektoni
The observer would know that the user wants to access the objective values. It then simply stores them. During training, a needs to be omitted to the observer, which then...
Nope, still lots of iterative algorithms/machines to port
Just checked a few machines that are iterative and realised that most low hanging fruits have been taken. There are quite a few algorithms to port left but the changes...