Hi Maryam Currently, no tree based methods have a partial fit method. We are currently working on expanding the tree module, you can see our checklist here; https://github.com/scikit-learn/scikit-learn/issues/5212
There are many methods to reduce the dimensionality of data, if you are using high dimensional data, including Gaussian random projections or LSH for continuous data, and collapsing samples into higher weight samples for discrete data. If you provide more information about your use case, I may be able to be of more help. Jacob On Wed, Sep 30, 2015 at 5:12 AM, Maryam Tavakol <maryam.tavakol...@gmail.com > wrote: > Dear all, > > I am using Gradient Boosting Classifier from scikit-learn for a huge set > of data. Unfortunately, the method loads the whole data into memory (around > 45 GBs!). As it is not very easy to modify the code to stream data, is > there any other way to make it scalable? > > Best Regards, > Maryam Tavakol > > > ------------------------------------------------------------------------------ > > _______________________________________________ > Scikit-learn-general mailing list > Scikit-learn-general@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/scikit-learn-general > >
------------------------------------------------------------------------------
_______________________________________________ Scikit-learn-general mailing list Scikit-learn-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
