Am 30.11.2012 17:39, schrieb Philipp Singer: > > I thought about building a custom transformer, that I can pass to the > pipeline that somehow appends the features to train and test data. But > the problem is, that I don't know exactly which data is used for the > splits (i.e., with sample). How would you do it with a list of tuples? > Hm, I'm not sure about the problem. The list of tuples would be X then, right. And you pass X to GridSearchCV. Let's assume for a moment that GridSearch can cope with list input (*) The it will use the cross-validation to generate indices test and train and just pass on X[train] to your transformer, where you could unpack it.
(*) Not sure if that actually happens - I think I messed this up in the past and now it resulted in issue #1137. Maybe we should just find my commit and revert it? In the meantime you can take arrays of dtype "object" instead of lists (maybe). Andy ------------------------------------------------------------------------------ Keep yourself connected to Go Parallel: TUNE You got it built. Now make it sing. Tune shows you how. http://goparallel.sourceforge.net _______________________________________________ Scikit-learn-general mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
