updated docs

--HG--
extra : convert_revision : svn%3Afdd8eb12-d10e-0410-9acb-85c331704f74/trunk%403665

docs/docs/ml.xml

@@ -579,23 +579,26 @@ Davis E. King. <a href="http://www.jmlr.org/papers/volume10/king09a/king09a.pdf"
          <description>
             <p>
                 This is an implementation of an online algorithm for recursively finding a
-                set of linearly independent vectors in a kernel induced feature space.  To 
-                use it you decide how large you would like the set to be and then you feed it 
-                sample points.  
+                set (aka dictionary) of linearly independent vectors in a kernel induced 
+                feature space.  To use it you decide how large you would like the dictionary 
+                to be and then you feed it sample points.  
             </p>
             <p>
-                
-                Each time you present it with a new sample point it either 
-                keeps the current set of independent points unchanged, or if the new point 
-                is "more linearly independent" than one of the points it already has,  
-                it replaces the weakly linearly independent point with the new one.
+                The implementation uses the Approximately Linearly Dependent metric described 
+                in the paper The Kernel Recursive Least Squares Algorithm by Yaakov Engel to 
+                decide which points are more linearly independent than others.  The metric is 
+                simply the squared distance between a test point and the subspace spanned by 
+                the current set of dictionary vectors.
             </p>
-
             <p>
-                
-                This object uses the Approximately Linearly Dependent metric described in the paper 
-                The Kernel Recursive Least Squares Algorithm by Yaakov Engel to decide which
-                points are more linearly independent than others.
+                Each time you present this object with a new sample point   
+                it calculates the projection distance and if it is sufficiently large then this 
+                new point is included into the current dictionary.  Note that this object can
+                be configured to have a maximum size.  Once the max dictionary size is reached
+                each new point kicks out a previous point.  This is done by selecting the current
+                dictionary vector that has the smallest projection error onto the others.  That 
+                is, the "least linearly independent" vector is removed to make room for the 
+                new one.
             </p>
          </description>
          <examples>