[mlpack-svn] [MLPACK] #299: Enhancement of PCA library
MLPACK Trac
trac at coffeetalk-1.cc.gatech.edu
Tue Aug 13 14:32:38 EDT 2013
#299: Enhancement of PCA library
---------------------------+------------------------------------------------
Reporter: sumedhghaisas | Owner:
Type: enhancement | Status: new
Priority: major | Milestone: mlpack 1.0.7
Component: mlpack | Keywords:
Blocking: | Blocked By:
---------------------------+------------------------------------------------
I have added some functions and modified some in pca.cpp and pca.hpp
Modification
1) Made use of svd for calculations which reduces many loops.
2) In the function Apply(data,newDimension) coeffs and eigVals were not
shredded. Fixed that.
Added functions -
1) Added the concept of variance retention.
2) Added function Apply(mat data,vec var_retained) which after
transforming the data return a vector var_retained. First entry of this
vector is variance retained if dimensions of current dataset is reduced to
1,and so on.
3) Added function Apply(mat data,size_t newDimension,double var_retained)
which return var_retained after reducing the dimensions of the data.
4) Added function Apply(mat data,double var_retained,size_t newDimension)
which reduces the dimension to minimum dimensions possible such that user
defined var_retained is maintained.
PFA both the files pca.cpp and pca.hpp
--
Ticket URL: <http://trac.research.cc.gatech.edu/fastlab/ticket/299>
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