When using filters in a pipeop it would be nice to have a general way of computing the (p)hash of the filter object. for most filters the class() is enough to differentiate what a filter does but for some this may not be the case, e.g. the importance filter (which filters depending on a learner given during construction).
When using filters in a pipeop it would be nice to have a general way of computing the (p)hash of the filter object. for most filters the
class()is enough to differentiate what a filter does but for some this may not be the case, e.g. the importance filter (which filters depending on a learner given during construction).