.TH "mlpack::nca::NCA< MetricType, OptimizerType >" 3 "Tue Sep 9 2014" "Version 1.0.10" "MLPACK" \" -*- nroff -*-
.ad l
.nh
.SH NAME
mlpack::nca::NCA< MetricType, OptimizerType > \- 
.PP
An implementation of Neighborhood Components Analysis, both a linear dimensionality reduction technique and a distance learning technique\&.  

.SH SYNOPSIS
.br
.PP
.SS "Public Member Functions"

.in +1c
.ti -1c
.RI "\fBNCA\fP (const arma::mat &\fBdataset\fP, const arma::Col< size_t > &\fBlabels\fP, MetricType \fBmetric\fP=MetricType())"
.br
.RI "\fIConstruct the Neighborhood Components Analysis object\&. \fP"
.ti -1c
.RI "const arma::mat & \fBDataset\fP () const "
.br
.RI "\fIGet the dataset reference\&. \fP"
.ti -1c
.RI "const arma::Col< size_t > & \fBLabels\fP () const "
.br
.RI "\fIGet the labels reference\&. \fP"
.ti -1c
.RI "void \fBLearnDistance\fP (arma::mat &outputMatrix)"
.br
.RI "\fIPerform Neighborhood Components Analysis\&. \fP"
.ti -1c
.RI "const OptimizerType
.br
< \fBSoftmaxErrorFunction\fP
.br
< MetricType > > & \fBOptimizer\fP () const "
.br
.RI "\fIGet the optimizer\&. \fP"
.ti -1c
.RI "OptimizerType
.br
< \fBSoftmaxErrorFunction\fP
.br
< MetricType > > & \fBOptimizer\fP ()"
.br
.ti -1c
.RI "std::string \fBToString\fP () const "
.br
.in -1c
.SS "Private Attributes"

.in +1c
.ti -1c
.RI "const arma::mat & \fBdataset\fP"
.br
.RI "\fIDataset reference\&. \fP"
.ti -1c
.RI "\fBSoftmaxErrorFunction\fP< MetricType > \fBerrorFunction\fP"
.br
.RI "\fIThe function to optimize\&. \fP"
.ti -1c
.RI "const arma::Col< size_t > & \fBlabels\fP"
.br
.RI "\fILabels reference\&. \fP"
.ti -1c
.RI "MetricType \fBmetric\fP"
.br
.RI "\fIMetric to be used\&. \fP"
.ti -1c
.RI "OptimizerType
.br
< \fBSoftmaxErrorFunction\fP
.br
< MetricType > > \fBoptimizer\fP"
.br
.RI "\fIThe optimizer to use\&. \fP"
.in -1c
.SH "Detailed Description"
.PP 

.SS "template<typename MetricType = metric::SquaredEuclideanDistance, template< typename > class OptimizerType = optimization::SGD>class mlpack::nca::NCA< MetricType, OptimizerType >"
An implementation of Neighborhood Components Analysis, both a linear dimensionality reduction technique and a distance learning technique\&. 

The method seeks to improve k-nearest-neighbor classification on a dataset by scaling the dimensions\&. The method is nonparametric, and does not require a value of k\&. It works by using stochastic ('soft') neighbor assignments and using optimization techniques over the gradient of the accuracy of the neighbor assignments\&.
.PP
For more details, see the following published paper:
.PP
.PP
.nf
@inproceedings{Goldberger2004,
  author = {Goldberger, Jacob and Roweis, Sam and Hinton, Geoff and
      Salakhutdinov, Ruslan},
  booktitle = {Advances in Neural Information Processing Systems 17},
  pages = {513--520},
  publisher = {MIT Press},
  title = {{Neighbourhood Components Analysis}},
  year = {2004}
}
.fi
.PP
 
.PP
Definition at line 59 of file nca\&.hpp\&.
.SH "Constructor & Destructor Documentation"
.PP 
.SS "template<typename MetricType  = metric::SquaredEuclideanDistance, template< typename > class OptimizerType = optimization::SGD> \fBmlpack::nca::NCA\fP< MetricType, OptimizerType >::\fBNCA\fP (const arma::mat &dataset, const arma::Col< size_t > &labels, MetricTypemetric = \fCMetricType()\fP)"

.PP
Construct the Neighborhood Components Analysis object\&. This simply stores the reference to the dataset and labels as well as the parameters for optimization before the actual optimization is performed\&.
.PP
\fBParameters:\fP
.RS 4
\fIdataset\fP Input dataset\&. 
.br
\fIlabels\fP Input dataset labels\&. 
.br
\fIstepSize\fP Step size for stochastic gradient descent\&. 
.br
\fImaxIterations\fP Maximum iterations for stochastic gradient descent\&. 
.br
\fItolerance\fP Tolerance for termination of stochastic gradient descent\&. 
.br
\fIshuffle\fP Whether or not to shuffle the dataset during SGD\&. 
.br
\fImetric\fP Instantiated metric to use\&. 
.RE
.PP

.SH "Member Function Documentation"
.PP 
.SS "template<typename MetricType  = metric::SquaredEuclideanDistance, template< typename > class OptimizerType = optimization::SGD> const arma::mat& \fBmlpack::nca::NCA\fP< MetricType, OptimizerType >::Dataset () const\fC [inline]\fP"

.PP
Get the dataset reference\&. 
.PP
Definition at line 91 of file nca\&.hpp\&.
.PP
References mlpack::nca::NCA< MetricType, OptimizerType >::dataset\&.
.SS "template<typename MetricType  = metric::SquaredEuclideanDistance, template< typename > class OptimizerType = optimization::SGD> const arma::Col<size_t>& \fBmlpack::nca::NCA\fP< MetricType, OptimizerType >::Labels () const\fC [inline]\fP"

.PP
Get the labels reference\&. 
.PP
Definition at line 93 of file nca\&.hpp\&.
.PP
References mlpack::nca::NCA< MetricType, OptimizerType >::labels\&.
.SS "template<typename MetricType  = metric::SquaredEuclideanDistance, template< typename > class OptimizerType = optimization::SGD> void \fBmlpack::nca::NCA\fP< MetricType, OptimizerType >::LearnDistance (arma::mat &outputMatrix)"

.PP
Perform Neighborhood Components Analysis\&. The output distance learning matrix is written into the passed reference\&. If \fBLearnDistance()\fP is called with an outputMatrix which has the correct size (dataset\&.n_rows x dataset\&.n_rows), that matrix will be used as the starting point for optimization\&.
.PP
\fBParameters:\fP
.RS 4
\fIoutput_matrix\fP Covariance matrix of Mahalanobis distance\&. 
.RE
.PP

.SS "template<typename MetricType  = metric::SquaredEuclideanDistance, template< typename > class OptimizerType = optimization::SGD> const OptimizerType<\fBSoftmaxErrorFunction\fP<MetricType> >& \fBmlpack::nca::NCA\fP< MetricType, OptimizerType >::Optimizer () const\fC [inline]\fP"

.PP
Get the optimizer\&. 
.PP
Definition at line 96 of file nca\&.hpp\&.
.PP
References mlpack::nca::NCA< MetricType, OptimizerType >::optimizer\&.
.SS "template<typename MetricType  = metric::SquaredEuclideanDistance, template< typename > class OptimizerType = optimization::SGD> OptimizerType<\fBSoftmaxErrorFunction\fP<MetricType> >& \fBmlpack::nca::NCA\fP< MetricType, OptimizerType >::Optimizer ()\fC [inline]\fP"

.PP
Definition at line 98 of file nca\&.hpp\&.
.PP
References mlpack::nca::NCA< MetricType, OptimizerType >::optimizer\&.
.SS "template<typename MetricType  = metric::SquaredEuclideanDistance, template< typename > class OptimizerType = optimization::SGD> std::string \fBmlpack::nca::NCA\fP< MetricType, OptimizerType >::ToString () const"

.SH "Member Data Documentation"
.PP 
.SS "template<typename MetricType  = metric::SquaredEuclideanDistance, template< typename > class OptimizerType = optimization::SGD> const arma::mat& \fBmlpack::nca::NCA\fP< MetricType, OptimizerType >::dataset\fC [private]\fP"

.PP
Dataset reference\&. 
.PP
Definition at line 106 of file nca\&.hpp\&.
.PP
Referenced by mlpack::nca::NCA< MetricType, OptimizerType >::Dataset()\&.
.SS "template<typename MetricType  = metric::SquaredEuclideanDistance, template< typename > class OptimizerType = optimization::SGD> \fBSoftmaxErrorFunction\fP<MetricType> \fBmlpack::nca::NCA\fP< MetricType, OptimizerType >::errorFunction\fC [private]\fP"

.PP
The function to optimize\&. 
.PP
Definition at line 114 of file nca\&.hpp\&.
.SS "template<typename MetricType  = metric::SquaredEuclideanDistance, template< typename > class OptimizerType = optimization::SGD> const arma::Col<size_t>& \fBmlpack::nca::NCA\fP< MetricType, OptimizerType >::labels\fC [private]\fP"

.PP
Labels reference\&. 
.PP
Definition at line 108 of file nca\&.hpp\&.
.PP
Referenced by mlpack::nca::NCA< MetricType, OptimizerType >::Labels()\&.
.SS "template<typename MetricType  = metric::SquaredEuclideanDistance, template< typename > class OptimizerType = optimization::SGD> MetricType \fBmlpack::nca::NCA\fP< MetricType, OptimizerType >::metric\fC [private]\fP"

.PP
Metric to be used\&. 
.PP
Definition at line 111 of file nca\&.hpp\&.
.SS "template<typename MetricType  = metric::SquaredEuclideanDistance, template< typename > class OptimizerType = optimization::SGD> OptimizerType<\fBSoftmaxErrorFunction\fP<MetricType> > \fBmlpack::nca::NCA\fP< MetricType, OptimizerType >::optimizer\fC [private]\fP"

.PP
The optimizer to use\&. 
.PP
Definition at line 117 of file nca\&.hpp\&.
.PP
Referenced by mlpack::nca::NCA< MetricType, OptimizerType >::Optimizer()\&.

.SH "Author"
.PP 
Generated automatically by Doxygen for MLPACK from the source code\&.