[mlpack-svn] [MLPACK] #370: Mean-shift clustering method

[MLPACK Trac]

#370: Mean-shift clustering method
-----------------------+----------------------------------------------------
  Reporter:  alexshtf  |        Owner:     
      Type:  defect    |       Status:  new
  Priority:  major     |    Milestone:     
 Component:  mlpack    |   Resolution:     
  Keywords:            |     Blocking:     
Blocked By:            |  
-----------------------+----------------------------------------------------

Comment (by rcurtin):

 Hi Alex,

 Sorry for the slow response.  An implementation of mean shift would be
 gladly accepted assuming that it adheres to the mlpack coding conventions
 (see the wiki page NewStyleGuidelines) and compares favorably or at least
 comparably (with respect to runtime) to other implementations (for
 instance, scikit-learn has an implementation).

 There is a fast mean shift algorithm using dual-tree algorithms, too:

 'Wang, P., Lee, D., Gray, A. G., & Rehg, J. M. (2007). Fast mean shift
 with accurate and stable convergence. In International Conference on
 Artificial Intelligence and Statistics (pp. 604-611).'
 http://machinelearning.wustl.edu/mlpapers/paper_files/AISTATS07_WangLGR.pdf

 But I only point that out as possible future work.  For now the typical
 algorithm will work.

 If you need help, feel free to either comment on this bug, come to #mlpack
 on freenode, or use the mailing list.  I'd be happy to help work out
 issues.

 Thanks,

 Ryan

-- 
Ticket URL: <http://trac.research.cc.gatech.edu/fastlab/ticket/370#comment:2>
MLPACK <www.fast-lab.org>
MLPACK is an intuitive, fast, and scalable C++ machine learning library developed at Georgia Tech.