Hey everybody. I'm happy to announce the release of scikit-learn 0.17. A big thank you to everybody who contributed.
Highlights of the release include: - Latent Dirichlet Allocation via Online Variational Inference - A Stochastic Average Gradient solver for LogisticRegression and RidgeRegression - A Coordinate Descent solver for NMF (which is much faster!) - The Barnes-Hut approximation for T-SNE (much faster!) - Many new cross-validation strategies and preprocessing transformers - Speed improvements to the tree-based models - and so much more... Read the full change-log here: http://scikit-learn.org/stable/whats_new.html An important API change is that we deprecated support for X being a 1d array. Please make sure that X is always (n_features, n_samples), even if one of them is one. The new version is available via pip and conda: $ pip install --update scikit-learn or $ conda update scikit-learn Please don't update a (ana)conda installation using pip, as that might lead to problems. Let us know any issues you have on the issue tracker: https://github.com/scikit-learn/scikit-learn/issues Happy learning! Andy ------------------------------------------------------------------------------ _______________________________________________ Scikit-learn-general mailing list Scikit-learn-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
