\section{S\+V\+D\+Batch\+Learning Class Reference}
\label{classmlpack_1_1amf_1_1SVDBatchLearning}\index{S\+V\+D\+Batch\+Learning@{S\+V\+D\+Batch\+Learning}}


This class implements S\+VD batch learning with momentum.  


\subsection*{Public Member Functions}
\begin{DoxyCompactItemize}
\item 
\textbf{ S\+V\+D\+Batch\+Learning} (double u=0.\+0002, double kw=0, double kh=0, double momentum=0.\+9)
\begin{DoxyCompactList}\small\item\em S\+VD Batch learning constructor. \end{DoxyCompactList}\item 
{\footnotesize template$<$typename Mat\+Type $>$ }\\void \textbf{ H\+Update} (const Mat\+Type \&V, const arma\+::mat \&W, arma\+::mat \&H)
\begin{DoxyCompactList}\small\item\em The update rule for the encoding matrix H. \end{DoxyCompactList}\item 
{\footnotesize template$<$typename Mat\+Type $>$ }\\void \textbf{ Initialize} (const Mat\+Type \&dataset, const size\+\_\+t rank)
\begin{DoxyCompactList}\small\item\em Initialize parameters before factorization. \end{DoxyCompactList}\item 
{\footnotesize template$<$typename Archive $>$ }\\void \textbf{ serialize} (Archive \&ar, const unsigned int)
\begin{DoxyCompactList}\small\item\em Serialize the S\+V\+D\+Batch object. \end{DoxyCompactList}\item 
{\footnotesize template$<$typename Mat\+Type $>$ }\\void \textbf{ W\+Update} (const Mat\+Type \&V, arma\+::mat \&W, const arma\+::mat \&H)
\begin{DoxyCompactList}\small\item\em The update rule for the basis matrix W. \end{DoxyCompactList}\end{DoxyCompactItemize}


\subsection{Detailed Description}
This class implements S\+VD batch learning with momentum. 

This procedure is described in the following paper\+:


\begin{DoxyCode}
@techreport\{ma2008guide,
  title=\{A Guide to Singular Value Decomposition \textcolor{keywordflow}{for} Collaborative
      Filtering\},
  author=\{Ma, Chih-Chao\},
  year=\{2008\},
  institution=\{Department of Computer Science, National Taiwan University\}
\}
\end{DoxyCode}


This class implements \textquotesingle{}Algorithm 4\textquotesingle{} as given in the paper.

The factorizer decomposes the matrix V into two matrices W and H such that sum of sum of squared error between V and W $\ast$ H is minimum. This optimization is performed with gradient descent. To make gradient descent faster, momentum is added. 

Definition at line 41 of file svd\+\_\+batch\+\_\+learning.\+hpp.



\subsection{Constructor \& Destructor Documentation}
\mbox{\label{classmlpack_1_1amf_1_1SVDBatchLearning_aaec0e3862d8b79920ea8ecda25dfcfba}} 
\index{mlpack\+::amf\+::\+S\+V\+D\+Batch\+Learning@{mlpack\+::amf\+::\+S\+V\+D\+Batch\+Learning}!S\+V\+D\+Batch\+Learning@{S\+V\+D\+Batch\+Learning}}
\index{S\+V\+D\+Batch\+Learning@{S\+V\+D\+Batch\+Learning}!mlpack\+::amf\+::\+S\+V\+D\+Batch\+Learning@{mlpack\+::amf\+::\+S\+V\+D\+Batch\+Learning}}
\subsubsection{S\+V\+D\+Batch\+Learning()}
{\footnotesize\ttfamily \textbf{ S\+V\+D\+Batch\+Learning} (\begin{DoxyParamCaption}\item[{double}]{u = {\ttfamily 0.0002},  }\item[{double}]{kw = {\ttfamily 0},  }\item[{double}]{kh = {\ttfamily 0},  }\item[{double}]{momentum = {\ttfamily 0.9} }\end{DoxyParamCaption})\hspace{0.3cm}{\ttfamily [inline]}}



S\+VD Batch learning constructor. 


\begin{DoxyParams}{Parameters}
{\em u} & step value used in batch learning \\
\hline
{\em kw} & regularization constant for W matrix \\
\hline
{\em kh} & regularization constant for H matrix \\
\hline
{\em momentum} & momentum applied to batch learning process \\
\hline
\end{DoxyParams}


Definition at line 52 of file svd\+\_\+batch\+\_\+learning.\+hpp.



\subsection{Member Function Documentation}
\mbox{\label{classmlpack_1_1amf_1_1SVDBatchLearning_a3c338e327964c027a7d4172e638a5dff}} 
\index{mlpack\+::amf\+::\+S\+V\+D\+Batch\+Learning@{mlpack\+::amf\+::\+S\+V\+D\+Batch\+Learning}!H\+Update@{H\+Update}}
\index{H\+Update@{H\+Update}!mlpack\+::amf\+::\+S\+V\+D\+Batch\+Learning@{mlpack\+::amf\+::\+S\+V\+D\+Batch\+Learning}}
\subsubsection{H\+Update()}
{\footnotesize\ttfamily void H\+Update (\begin{DoxyParamCaption}\item[{const Mat\+Type \&}]{V,  }\item[{const arma\+::mat \&}]{W,  }\item[{arma\+::mat \&}]{H }\end{DoxyParamCaption})\hspace{0.3cm}{\ttfamily [inline]}}



The update rule for the encoding matrix H. 

The function takes in all the matrices and only changes the value of the H matrix.


\begin{DoxyParams}{Parameters}
{\em V} & Input matrix to be factorized. \\
\hline
{\em W} & Basis matrix. \\
\hline
{\em H} & Encoding matrix to be updated. \\
\hline
\end{DoxyParams}


Definition at line 133 of file svd\+\_\+batch\+\_\+learning.\+hpp.

\mbox{\label{classmlpack_1_1amf_1_1SVDBatchLearning_a1af73fa03a54127db8d2bbc558a58c4c}} 
\index{mlpack\+::amf\+::\+S\+V\+D\+Batch\+Learning@{mlpack\+::amf\+::\+S\+V\+D\+Batch\+Learning}!Initialize@{Initialize}}
\index{Initialize@{Initialize}!mlpack\+::amf\+::\+S\+V\+D\+Batch\+Learning@{mlpack\+::amf\+::\+S\+V\+D\+Batch\+Learning}}
\subsubsection{Initialize()}
{\footnotesize\ttfamily void Initialize (\begin{DoxyParamCaption}\item[{const Mat\+Type \&}]{dataset,  }\item[{const size\+\_\+t}]{rank }\end{DoxyParamCaption})\hspace{0.3cm}{\ttfamily [inline]}}



Initialize parameters before factorization. 

This function must be called before a new factorization. This resets the internally-\/held momentum.


\begin{DoxyParams}{Parameters}
{\em dataset} & Input matrix to be factorized. \\
\hline
{\em rank} & rank of factorization \\
\hline
\end{DoxyParams}


Definition at line 69 of file svd\+\_\+batch\+\_\+learning.\+hpp.

\mbox{\label{classmlpack_1_1amf_1_1SVDBatchLearning_af0dd9205158ccf7bcfcd8ff81f79c927}} 
\index{mlpack\+::amf\+::\+S\+V\+D\+Batch\+Learning@{mlpack\+::amf\+::\+S\+V\+D\+Batch\+Learning}!serialize@{serialize}}
\index{serialize@{serialize}!mlpack\+::amf\+::\+S\+V\+D\+Batch\+Learning@{mlpack\+::amf\+::\+S\+V\+D\+Batch\+Learning}}
\subsubsection{serialize()}
{\footnotesize\ttfamily void serialize (\begin{DoxyParamCaption}\item[{Archive \&}]{ar,  }\item[{const unsigned}]{int }\end{DoxyParamCaption})\hspace{0.3cm}{\ttfamily [inline]}}



Serialize the S\+V\+D\+Batch object. 



Definition at line 169 of file svd\+\_\+batch\+\_\+learning.\+hpp.

\mbox{\label{classmlpack_1_1amf_1_1SVDBatchLearning_a2df8bfa574f8a8e43e6b1466dc0526fe}} 
\index{mlpack\+::amf\+::\+S\+V\+D\+Batch\+Learning@{mlpack\+::amf\+::\+S\+V\+D\+Batch\+Learning}!W\+Update@{W\+Update}}
\index{W\+Update@{W\+Update}!mlpack\+::amf\+::\+S\+V\+D\+Batch\+Learning@{mlpack\+::amf\+::\+S\+V\+D\+Batch\+Learning}}
\subsubsection{W\+Update()}
{\footnotesize\ttfamily void W\+Update (\begin{DoxyParamCaption}\item[{const Mat\+Type \&}]{V,  }\item[{arma\+::mat \&}]{W,  }\item[{const arma\+::mat \&}]{H }\end{DoxyParamCaption})\hspace{0.3cm}{\ttfamily [inline]}}



The update rule for the basis matrix W. 

The function takes in all the matrices and only changes the value of the W matrix.


\begin{DoxyParams}{Parameters}
{\em V} & Input matrix to be factorized. \\
\hline
{\em W} & Basis matrix to be updated. \\
\hline
{\em H} & Encoding matrix. \\
\hline
\end{DoxyParams}


Definition at line 88 of file svd\+\_\+batch\+\_\+learning.\+hpp.



The documentation for this class was generated from the following file\+:\begin{DoxyCompactItemize}
\item 
/var/www/mlpack.\+ratml.\+org/mlpack.\+org/\+\_\+src/mlpack-\/3.\+3.\+2/src/mlpack/methods/amf/update\+\_\+rules/\textbf{ svd\+\_\+batch\+\_\+learning.\+hpp}\end{DoxyCompactItemize}