\section{Bias\+S\+VD$<$ Optimizer\+Type $>$ Class Template Reference}
\label{classmlpack_1_1svd_1_1BiasSVD}\index{Bias\+S\+V\+D$<$ Optimizer\+Type $>$@{Bias\+S\+V\+D$<$ Optimizer\+Type $>$}}


Bias S\+VD is an improvement on Regularized S\+VD which is a matrix factorization techniques.  


\subsection*{Public Member Functions}
\begin{DoxyCompactItemize}
\item 
\textbf{ Bias\+S\+VD} (const size\+\_\+t iterations=10, const double alpha=0.\+02, const double lambda=0.\+05)
\begin{DoxyCompactList}\small\item\em Constructor of Bias S\+VD. \end{DoxyCompactList}\item 
void \textbf{ Apply} (const arma\+::mat \&data, const size\+\_\+t rank, arma\+::mat \&u, arma\+::mat \&v, arma\+::vec \&p, arma\+::vec \&q)
\begin{DoxyCompactList}\small\item\em Trains the model and obtains user/item matrices and user/item bias. \end{DoxyCompactList}\end{DoxyCompactItemize}


\subsection{Detailed Description}
\subsubsection*{template$<$typename Optimizer\+Type = ens\+::\+Standard\+S\+GD$>$\newline
class mlpack\+::svd\+::\+Bias\+S\+V\+D$<$ Optimizer\+Type $>$}

Bias S\+VD is an improvement on Regularized S\+VD which is a matrix factorization techniques. 

Bias S\+VD outputs user/item latent vectors and user/item bias, so that $ r_{iu} = b_i + b_u + p_i * q_u $, where b, p, q are bias, item latent, user latent respectively. Parameters are optmized by Stochastic Gradient Desent(\+S\+G\+D). The updates also penalize the learning of large feature values by means of regularization.

An example of how to use the interface is shown below\+:


\begin{DoxyCode}
arma::mat data; \textcolor{comment}{// Rating data in the form of coordinate list.}

\textcolor{keyword}{const} \textcolor{keywordtype}{size\_t} rank = 10; \textcolor{comment}{// Rank used for the decomposition.}
\textcolor{keyword}{const} \textcolor{keywordtype}{size\_t} iterations = 10; \textcolor{comment}{// Number of iterations used for optimization.}

\textcolor{keyword}{const} \textcolor{keywordtype}{double} alpha = 0.005 \textcolor{comment}{// Learning rate for the SGD optimizer.}
\textcolor{keyword}{const} \textcolor{keywordtype}{double} lambda = 0.02 \textcolor{comment}{// Regularization parameter for the optimization.}

\textcolor{comment}{// Make a BiasSVD object.}
BiasSVD<> biasSVD(iterations, alpha, lambda);

arma::mat u, v; \textcolor{comment}{// Item and User matrices.}
arma::vec p, q; \textcolor{comment}{// Item and User bias.}

\textcolor{comment}{// Use the Apply() method to get a factorization.}
rSVD.Apply(data, rank, u, v, p, q);
\end{DoxyCode}
 

Definition at line 57 of file bias\+\_\+svd.\+hpp.



\subsection{Constructor \& Destructor Documentation}
\mbox{\label{classmlpack_1_1svd_1_1BiasSVD_a217884b3b86ec1fa9148eac852165945}} 
\index{mlpack\+::svd\+::\+Bias\+S\+VD@{mlpack\+::svd\+::\+Bias\+S\+VD}!Bias\+S\+VD@{Bias\+S\+VD}}
\index{Bias\+S\+VD@{Bias\+S\+VD}!mlpack\+::svd\+::\+Bias\+S\+VD@{mlpack\+::svd\+::\+Bias\+S\+VD}}
\subsubsection{Bias\+S\+V\+D()}
{\footnotesize\ttfamily \textbf{ Bias\+S\+VD} (\begin{DoxyParamCaption}\item[{const size\+\_\+t}]{iterations = {\ttfamily 10},  }\item[{const double}]{alpha = {\ttfamily 0.02},  }\item[{const double}]{lambda = {\ttfamily 0.05} }\end{DoxyParamCaption})}



Constructor of Bias S\+VD. 

By default S\+GD optimizer is used in \doxyref{Bias\+S\+VD}{p.}{classmlpack_1_1svd_1_1BiasSVD}. The optimizer uses a template specialization of Optimize().


\begin{DoxyParams}{Parameters}
{\em iterations} & Number of optimization iterations. \\
\hline
{\em alpha} & Learning rate for the S\+GD optimizer. \\
\hline
{\em lambda} & Regularization parameter for the optimization. \\
\hline
\end{DoxyParams}


\subsection{Member Function Documentation}
\mbox{\label{classmlpack_1_1svd_1_1BiasSVD_a2cda95c3460611ea7af141f8db7fb3eb}} 
\index{mlpack\+::svd\+::\+Bias\+S\+VD@{mlpack\+::svd\+::\+Bias\+S\+VD}!Apply@{Apply}}
\index{Apply@{Apply}!mlpack\+::svd\+::\+Bias\+S\+VD@{mlpack\+::svd\+::\+Bias\+S\+VD}}
\subsubsection{Apply()}
{\footnotesize\ttfamily void Apply (\begin{DoxyParamCaption}\item[{const arma\+::mat \&}]{data,  }\item[{const size\+\_\+t}]{rank,  }\item[{arma\+::mat \&}]{u,  }\item[{arma\+::mat \&}]{v,  }\item[{arma\+::vec \&}]{p,  }\item[{arma\+::vec \&}]{q }\end{DoxyParamCaption})}



Trains the model and obtains user/item matrices and user/item bias. 


\begin{DoxyParams}{Parameters}
{\em data} & Rating data matrix. \\
\hline
{\em rank} & Rank parameter to be used for optimization. \\
\hline
{\em u} & Item matrix obtained on decomposition. \\
\hline
{\em v} & User matrix obtained on decomposition. \\
\hline
{\em p} & Item bias. \\
\hline
{\em q} & User bias. \\
\hline
\end{DoxyParams}


Referenced by Bias\+S\+V\+D\+Policy\+::\+Apply().



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.\+1/src/mlpack/methods/bias\+\_\+svd/\textbf{ bias\+\_\+svd.\+hpp}\end{DoxyCompactItemize}