May 15, 2008 - Soeren Sonnenburg - email@example.com
SHOGUN Release version 0.6.2 (libshogun 0.0,
data 0.0, parameter 0.0)
Experimental support for r-modular thanks to the swig team!
All python-modular examples describing the use of kernels, classifier, distributions, features, distances, regression and preprocessors have been ported to r-modular (requires swig from svn).
The 'send_command' legacy is no longer necessary, numbers can be used as such and don't have to be given as strings. All examples for r,python,octave,matlab have been converted to the new syntax.
Resurrected the command line interface. Basic functionionality, like training a classifier on strings/real valued (sparse) features was restored. Readline completion was added.
- The weighted spectrum kernel is now working again.
- Fix the testsuite to reliably check methods that use random().
- Off-by-one bug fix in FixedDegreeStringKernel.
- Fix reading strings from file, when strings don't have the same length.
- Fix massive slowdown in modular interfaces for WD based kernels (it is 5-30 times faster now).