[mlpack-svn] [MLPACK] #344: Using Welford method to calculate variance in Naive Bayes Classifier

MLPACK Trac trac at coffeetalk-1.cc.gatech.edu
Wed Apr 16 14:24:30 EDT 2014


#344: Using Welford method to calculate variance in Naive Bayes Classifier
-----------------------------------+----------------------------------------
  Reporter:  akvah                 |        Owner:  rcurtin     
      Type:  enhancement           |       Status:  closed      
  Priority:  minor                 |    Milestone:  mlpack 1.0.9
 Component:  mlpack                |   Resolution:  fixed       
  Keywords:  variance calculation  |     Blocking:              
Blocked By:                        |  
-----------------------------------+----------------------------------------

Comment (by rcurtin):

 Ah, I see what you mean now.  I've committed the two-pass algorithm in
 r16430, and added a note that the naive algorithm could potentially be
 faster, which may be a useful comment if later we find that the mlpack
 implementation are slower than other implementations.  Currently the
 benchmarks show mlpack to be way slower for the naive Bayes classifier,
 but they need to be re-run; your opening this ticket made me look into why
 it was so slow and I committed some other fixes that will give drastic
 speed gains (r16426).  Thanks again for the contribution, and feel free to
 reopen the bug if I've missed something again. :)

-- 
Ticket URL: <http://trac.research.cc.gatech.edu/fastlab/ticket/344#comment:6>
MLPACK <www.fast-lab.org>
MLPACK is an intuitive, fast, and scalable C++ machine learning library developed at Georgia Tech.


More information about the mlpack-svn mailing list