Hi, thanks for the release!
> - Do not derive NaN scaling for SVM's C whenever data is > degenerate (lead to never finishing SVM training). Is this only relevant for automatic scaling of value according to the norm of the data (C=-1)? is this problem 'always' leading to never finished SVM training, or do you suggest to re-run previous (finished) analysis? Thanks, Matthias PyMVPA Team wrote: > Dear PyMVPA users, > > I am glad to announce that we are releasing 0.4.4 version of > PyMVPA. > > So far we have: > * pushed/tagged sources in git repository > (git://git.debian.org/git/pkg-exppsy/pymvpa.git) > * uploaded Debian packages into Debian/sid proper > (should become available later on today) > * uploaded Debian packages into http://neuro.debian.org > repository for: > + Debian: lenny, squeeze > + Ubuntu: jaunty, karmic > (http://neuro.debian.net/pkgs/python-mvpa.html) > * source tarballs and windows installer available from > usual location > https://alioth.debian.org/frs/?group_id=30954 > * RPM-based GNU/Linux Distributions > available from OpenSUSE Build Service. See > http://pymvpa.org/installation.html#rpm-based-gnu-linux-distributions > > This is primarily a bugfix release, probably the last in 0.4 series since > development for 0.5 release is leaping forward -- please keep an eye on the > mailing list for the announcement of availability of snapshots of development > version. > > > Changelog for the release is: > > * New functionality (19 NF commits): > > - :class:`~mvpa.clfs.gnb.GNB` implements Gaussian Naïve Bayes > Classifier. > - :func:`~mvpa.misc.fsl.base.read_fsl_design` to read FSL FEAT design.fsf > files (Contributed by Russell A. Poldrack). > - :class:`~mvpa.datasets.miscfx.SequenceStats` to provide basic > statistics on labels sequence (counter-balancing, > autocorrelation). > - New exceptions :class:`~mvpa.clfs.base.DegenerateInputError` and > :class:`~mvpa.clfs.base.FailedToTrainError` to be thrown by > classifiers primarily during training/testing. > - Debug target `STATMC` to report on progress of Monte-Carlo > sampling (during permutation testing). > > * Refactored (15 RF commits): > > - To get users prepared to 0.5 release, internally and in some > examples/documentation, access to states and > parameters is done via corresponding collections, not from the > top level object (e.g. `clf.states.predictions` instead of > soon-to-be-deprecated `clf.predictions`). That should lead also > to improved performance. > - Adopted copy.py from python2.6 (support Ellipsis as well). > > * Fixed (38 BF commits): > > - GLM output does not depend on the enabled states any more. > - Variety of docstrings fixed and/or improved. > - Do not derive NaN scaling for SVM's C whenever data is > degenerate (lead to never finishing SVM training). > - :mod:`~mvpa.clfs.sg` : > > + KRR is optional now -- avoids crashing if KRR is not available. > + tolerance to absent `set_precompute_matrix` in svmlight in > recent shogun versions. > + support for recent (present in 0.9.1) API change in exposing > debug levels. > > - Python 2.4 compatibility issues: :class:`~mvpa.clfs.knn.kNN` and > :class:`~mvpa.featsel.ifs.IFS` > > Enjoy! > > > ------------------------------------------------------------------------ > > _______________________________________________ > Pkg-ExpPsy-PyMVPA mailing list > [email protected] > http://lists.alioth.debian.org/mailman/listinfo/pkg-exppsy-pymvpa _______________________________________________ Pkg-ExpPsy-PyMVPA mailing list [email protected] http://lists.alioth.debian.org/mailman/listinfo/pkg-exppsy-pymvpa
