algorithmrvm.html,v

包含了模式识别中常用的一些分类器设计算法

<!-- ============ METHOD DETAIL ========== -->

<A NAME="method_detail"><!-- --></A>
<TABLE BORDER="1" WIDTH="100%" CELLPADDING="3" CELLSPACING="0" SUMMARY="">
<TR BGCOLOR="#CCCCFF" CLASS="TableHeadingColor">
<TH ALIGN="left" COLSPAN="1"><FONT SIZE="+2">
<B>Method Detail</B></FONT></TH>
</TR>
</TABLE>

<A NAME="initialize()"><!-- --></A><H3>
initialize</H3>
<PRE>
public boolean <B>initialize</B>()</PRE>
<DL>
<DD>Overrides the initialize() method in the base class.  Initializes
 member data and prepares for execution of first step.  This method
 "resets" the algorithm.
<P>
<DD><DL>
<DT><B>Specified by:</B><DD><CODE><A HREF="Algorithm.html#initialize()">initialize</A></CODE> in class <CODE><A HREF="Algorithm.html" title="class in &lt;Unnamed&gt;">Algorithm</A></CODE></DL>
</DD>
<DD><DL>

<DT><B>Returns:</B><DD>true</DL>
</DD>
</DL>
<HR>

<A NAME="run()"><!-- --></A><H3>
run</H3>
<PRE>
public void <B>run</B>()</PRE>
<DL>
<DD>Implementation of the run function from the Runnable interface.
 Determines what the current step is and calls the appropriate method.
<P>
<DD><DL>
<DT><B>Specified by:</B><DD><CODE>run</CODE> in interface <CODE>java.lang.Runnable</CODE><DT><B>Specified by:</B><DD><CODE><A HREF="Algorithm.html#run()">run</A></CODE> in class <CODE><A HREF="Algorithm.html" title="class in &lt;Unnamed&gt;">Algorithm</A></CODE></DL>
</DD>
<DD><DL>
</DL>
</DD>
</DL>
<HR>

<A NAME="trainFull()"><!-- --></A><H3>
trainFull</H3>
<PRE>
public boolean <B>trainFull</B>()</PRE>
<DL>
<DD>this method trains an RVM probabilistic classifier on the input data and
 targets provided. the inputs and targets should have a one-to-one
 correspondence and all targets should be either 0 (out-of-class) or
 1 (in-class). The training scheme follows that of [1] section 3.
 It is assumed that the class data and targets are already set when
 this method is called.

 the training algorithm in [1] for RVMs proceeds in three iterative steps

   1. prune away any weights whose hyperparameters have gone to infinity

   2. estimate most probable weights: in this step we find those
   weights that maximize equation (24) of [1].  The iteratively
   reweighted least squares algorithm is used to find w_mp

   3. estimate the covariance of a Gaussian approximation to the
   posterior distribution (the posterior is what we want to
   model in the end) over the weights centered at the weights,
   w_mp.

   4. estimate the hyperparameters that govern the weights. this
   is done by evaluating the derivative over the hyperparameters
   and finding the maximizing hyperparameters.

 1, 2, 3, and 4 are carried out iteratively until a suitable convergence
 criteria is satisfied.
<P>
<DD><DL>

<DT><B>Returns:</B><DD>boolean value indicating status</DL>
</DD>
</DL>
<HR>

<A NAME="initFullTrain()"><!-- --></A><H3>
initFullTrain</H3>
<PRE>
public boolean <B>initFullTrain</B>()</PRE>
<DL>
<DD>initializes the data structures in preparation for a full training
 pass
<P>
<DD><DL>

<DT><B>Returns:</B><DD>boolean value indicating status</DL>
</DD>
</DL>
<HR>

<A NAME="updateHyperparametersFull()"><!-- --></A><H3>
updateHyperparametersFull</H3>
<PRE>
public boolean <B>updateHyperparametersFull</B>()</PRE>
<DL>
<DD>Updates the hyperparameter values
<P>
<DD><DL>

<DT><B>Returns:</B><DD>boolean value indicating status</DL>
</DD>
</DL>
<HR>

<A NAME="irlsTrain()"><!-- --></A><H3>
irlsTrain</H3>
<PRE>
public boolean <B>irlsTrain</B>()</PRE>
<DL>
<DD>Completes one pass of IRLS training to update the weights
 given the currently assigne hyperparamters. In this step, we find
 those weights that maximize equation (24) of [1].  The iteratively
 reweighted least squares algorithm is used to find w_mp. It proceeds
 as follows:

   a. initialize
   b. loop until convergence
      1. compute the B matrix
      2. compute the Hessian and the gradient
      3. update weights with formula
          new weights = weights - inv(Hessian) * gradient
      4. check convergence
<P>
<DD><DL>

<DT><B>Returns:</B><DD>boolean value indicating status</DL>
</DD>
</DL>
<HR>

<A NAME="pruneAndUpdate()"><!-- --></A><H3>
pruneAndUpdate</H3>
<PRE>
public boolean <B>pruneAndUpdate</B>()</PRE>
<DL>
<DD>prunes off vectors whose hyperparameters have gone to infinity and
 updates working data sets
<P>
<DD><DL>

<DT><B>Returns:</B><DD>true</DL>
</DD>
</DL>
<HR>

<A NAME="pruneWeights()"><!-- --></A><H3>
pruneWeights</H3>
<PRE>
public boolean <B>pruneWeights</B>()</PRE>
<DL>
<DD>Prunes off vectors which attain a zero weight during
 training. Auxiliary data structures that are indexed according to the
 working set are also updated.
<P>
<DD><DL>

<DT><B>Returns:</B><DD>true</DL>
</DD>
</DL>
<HR>

<A NAME="computeVarianceCholesky()"><!-- --></A><H3>
computeVarianceCholesky</H3>
<PRE>
public boolean <B>computeVarianceCholesky</B>()</PRE>
<DL>
<DD>Computes the diagonal elements of the inverse from the
 cholesky decomposition
<P>
<DD><DL>

<DT><B>Returns:</B><DD>true</DL>
</DD>
</DL>
<HR>

<A NAME="computeSigma()"><!-- --></A><H3>
computeSigma</H3>
<PRE>
public boolean <B>computeSigma</B>()</PRE>
<DL>
<DD>Computes the sigma valued defined on pp. 218 of [1].
<P>
<DD><DL>

<DT><B>Returns:</B><DD>true</DL>
</DD>
</DL>
<HR>

<A NAME="computeLikelihood()"><!-- --></A><H3>
computeLikelihood</H3>
<PRE>
public double <B>computeLikelihood</B>()</PRE>
<DL>
<DD>Computes the log likelihood of the weights given the
 data. We would expect this value to increase as training proceeds
<P>
<DD><DL>

<DT><B>Returns:</B><DD>Likelihood of weights given the data</DL>
</DD>
</DL>
<HR>

<A NAME="computeDecisionRegions()"><!-- --></A><H3>
computeDecisionRegions</H3>
<PRE>
public void <B>computeDecisionRegions</B>()</PRE>
<DL>
<DD>Computes the line of discrimination for the classification 
 algorithms when the corresponding flags have been initialized
<P>
<DD><DL>
</DL>
</DD>
</DL>
<HR>

<A NAME="computeErrors()"><!-- --></A><H3>
computeErrors</H3>
<PRE>
public void <B>computeErrors</B>()</PRE>
<DL>
<DD>Computes the classification error for the data points
<P>
<DD><DL>
</DL>
</DD>
</DL>
<!-- ========= END OF CLASS DATA ========= -->
<HR>


<!-- ======= START OF BOTTOM NAVBAR ====== -->
<A NAME="navbar_bottom"><!-- --></A>
<A HREF="#skip-navbar_bottom" title="Skip navigation links"></A>
<TABLE BORDER="0" WIDTH="100%" CELLPADDING="1" CELLSPACING="0" SUMMARY="">
<TR>
<TD COLSPAN=2 BGCOLOR="#EEEEFF" CLASS="NavBarCell1">
<A NAME="navbar_bottom_firstrow"><!-- --></A>
<TABLE BORDER="0" CELLPADDING="0" CELLSPACING="3" SUMMARY="">
  <TR ALIGN="center" VALIGN="top">
  <TD BGCOLOR="#EEEEFF" CLASS="NavBarCell1">    <A HREF="package-summary.html"><FONT CLASS="NavBarFont1"><B>Package</B></FONT></A>&nbsp;</TD>
  <TD BGCOLOR="#FFFFFF" CLASS="NavBarCell1Rev"> &nbsp;<FONT CLASS="NavBarFont1Rev"><B>Class</B></FONT>&nbsp;</TD>
  <TD BGCOLOR="#EEEEFF" CLASS="NavBarCell1">    <A HREF="package-tree.html"><FONT CLASS="NavBarFont1"><B>Tree</B></FONT></A>&nbsp;</TD>
  <TD BGCOLOR="#EEEEFF" CLASS="NavBarCell1">    <A HREF="deprecated-list.html"><FONT CLASS="NavBarFont1"><B>Deprecated</B></FONT></A>&nbsp;</TD>
  <TD BGCOLOR="#EEEEFF" CLASS="NavBarCell1">    <A HREF="index-all.html"><FONT CLASS="NavBarFont1"><B>Index</B></FONT></A>&nbsp;</TD>
  <TD BGCOLOR="#EEEEFF" CLASS="NavBarCell1">    <A HREF="help-doc.html"><FONT CLASS="NavBarFont1"><B>Help</B></FONT></A>&nbsp;</TD>
  </TR>
</TABLE>
</TD>
<TD ALIGN="right" VALIGN="top" ROWSPAN=3><EM>
</EM>
</TD>
</TR>

<TR>
<TD BGCOLOR="white" CLASS="NavBarCell2"><FONT SIZE="-2">
&nbsp;<A HREF="AlgorithmPF.html" title="class in &lt;Unnamed&gt;"><B>PREV CLASS</B></A>&nbsp;
&nbsp;<A HREF="AlgorithmSVM.html" title="class in &lt;Unnamed&gt;"><B>NEXT CLASS</B></A></FONT></TD>
<TD BGCOLOR="white" CLASS="NavBarCell2"><FONT SIZE="-2">
  <A HREF="index.html?AlgorithmRVM.html" target="_top"><B>FRAMES</B></A>  &nbsp;
&nbsp;<A HREF="AlgorithmRVM.html" target="_top"><B>NO FRAMES</B></A>  &nbsp;
&nbsp;<SCRIPT type="text/javascript">
  <!--
  if(window==top) {
    document.writeln('<A HREF="allclasses-noframe.html"><B>All Classes</B></A>');
  }
  //-->
</SCRIPT>
<NOSCRIPT>
  <A HREF="allclasses-noframe.html"><B>All Classes</B></A>
</NOSCRIPT>


</FONT></TD>
</TR>
<TR>
<TD VALIGN="top" CLASS="NavBarCell3"><FONT SIZE="-2">
  SUMMARY:&nbsp;NESTED&nbsp;|&nbsp;FIELD&nbsp;|&nbsp;<A HREF="#constructor_summary">CONSTR</A>&nbsp;|&nbsp;<A HREF="#method_summary">METHOD</A></FONT></TD>
<TD VALIGN="top" CLASS="NavBarCell3"><FONT SIZE="-2">
DETAIL:&nbsp;FIELD&nbsp;|&nbsp;<A HREF="#constructor_detail">CONSTR</A>&nbsp;|&nbsp;<A HREF="#method_detail">METHOD</A></FONT></TD>
</TR>
</TABLE>
<A NAME="skip-navbar_bottom"></A>
<!-- ======== END OF BOTTOM NAVBAR ======= -->

<HR>

</BODY>
</HTML>
@