add logger functionality to website

docs/functionality.html

@@ -176,6 +176,61 @@ <h2>Loss</h2>
       </tr>
     </table>
 
+    <h2>Logger</h2>
+
+    <table>
+      <tr>
+        <th>Logger</th>
+        <th>Description</th>
+        <th>Arguments</th>
+      </tr>
+      <tr>
+        <td><code>LoggerIteration</code></td>
+        <td>
+          The basic logger which loggs the actual iteration. This one is defined automatically by calling, e.g., <code>train(1000)</code>, as stopper with a maximal number of iteration as passed through <code>train()</code>.
+        </td>
+        <td>
+          <dt><b>use.as.stopper</b></dt> <dd>Boolean to indicate whether this logger should also be used as stopper.</dd>
+          <dt><b>max.iterations</b></dt> <dd>Integer value specifying the maximal numbers of iterations if the logger is used as stopper, otherwise this argument does not have any effect.</dd>
+        </td>
+      </tr>
+      <tr>
+        <td><code>LoggerInbagRisk</code></td>
+        <td>
+          This logger takes a <code>Loss</code> object and calculates the risk using the train dataset.
+        </td>
+        <td>
+          <dt><b>use.as.stopper</b></dt> <dd>Boolean to indicate whether this logger should also be used as stopper.</dd>
+          <dt><b>used.loss</b></dt> <dd><code>Loss</code> object which is used to calculate the risk. Note that this could also be a custom loss and therefore be used to track performance measures.</dd>
+          <dt><b>eps.for.break</b></dt> <dd>Numeric value indicating the relative improvement which is used to stop the algorithm if the true improvement falls under <code>eps.for.break</code>.</dd>
+        </td>
+      </tr>      
+      <tr>
+        <td><code>LoggerOobRisk</code></td>
+        <td>
+          This logger takes a <code>Loss</code> object and calculates the risk using a custom out of bag dataset.
+        </td>
+        <td>
+          <dt><b>use.as.stopper</b></dt> <dd>Boolean to indicate whether this logger should also be used as stopper.</dd>
+          <dt><b>used.loss</b></dt> <dd><code>Loss</code> object which is used to calculate the risk. Note that this could also be a custom loss and therefore be used to track performance measures.</dd>
+          <dt><b>eps.for.break</b></dt> <dd>Numeric value indicating the relative improvement which is used to stop the algorithm if the true improvement falls under <code>eps.for.break</code>.</dd>
+          <dt><b>oob.data</b></dt> <dd>List of <code>Data</code> objects which corresponds to the same data objects used in <code>addBaselearner()</code>. Note that each <code>Compboost</code> object has a member function <code>prepareData()</code> which automatically converts a <code>data.frame</code> to a list required by the logger.</dd>
+          <dt><b>oob.response</b></dt> <dd>Vector of the response values corresonding to <code>oob.data</code>.</dd>
+        </td>
+      </tr>
+      <tr>
+        <td><code>LoggerTime</code></td>
+        <td>
+          This logger can be used to measure the ellapsed time at each iteration.
+        </td>
+        <td>
+          <dt><b>use.as.stopper</b></dt> <dd>Boolean to indicate whether this logger should also be used as stopper.</dd>
+          <dt><b>max.time</b></dt> <dd>Integer value that indicates how much time the algorithm has for the training.</dd>
+          <dt><b>time.unit</b></dt> <dd>Character to specify the time unit. Possible choices are minutes, seconds or microseconds.</dd>
+        </td>
+      </tr>
+    </table>
+
 
 <!--
   <h2>Optimizer</h2>