Added some comments that document how the tolerance parameters relate to

projection error as given by the empirical_kernel_map.

--HG--
extra : convert_revision : svn%3Afdd8eb12-d10e-0410-9acb-85c331704f74/trunk%403408

dlib/svm/kcentroid_abstract.h

@@ -147,6 +147,14 @@ namespace dlib
                   estimate but also in less dictionary vectors.  (Note that in any case, 
                   the max_dictionary_size() limits the number of dictionary vectors no 
                   matter the setting of the tolerance)
+                - The exact meaning of the tolerance parameter is the following: 
+                  Imagine that we have an empirical_kernel_map that contains all
+                  the current dictionary vectors.  Then the tolerance is the minimum
+                  projection error (as given by empirical_kernel_map::project()) required
+                  to cause us to include a new vector in the dictionary.  So each time
+                  you call train() the kcentroid basically just computes the projection
+                  error for that new sample and if it is larger than the tolerance
+                  then that new sample becomes part of the dictionary.
         !*/
 
         void clear_dictionary (

dlib/svm/krls_abstract.h

@@ -75,6 +75,14 @@ namespace dlib
                   estimate while also resulting in a bigger set of dictionary vectors in 
                   the learned decision function.  Bigger tolerances values result in a 
                   less accurate decision function but also in less dictionary vectors.
+                - The exact meaning of the tolerance parameter is the following: 
+                  Imagine that we have an empirical_kernel_map that contains all
+                  the current dictionary vectors.  Then the tolerance is the minimum
+                  projection error (as given by empirical_kernel_map::project()) required
+                  to cause us to include a new vector in the dictionary.  So each time
+                  you call train() the krls object basically just computes the projection
+                  error for that new sample and if it is larger than the tolerance
+                  then that new sample becomes part of the dictionary.
         !*/
 
         const kernel_type& get_kernel (

dlib/svm/linearly_independent_subset_finder_abstract.h

@@ -95,6 +95,14 @@ namespace dlib
                   a sample must be for it to even be considered for addition to the dictionary.  
                   Moreover, bigger values of this field will make the algorithm run faster but 
                   might give less accurate results.
+                - The exact meaning of the tolerance parameter is the following: 
+                  Imagine that we have an empirical_kernel_map that contains all the current 
+                  dictionary vectors.  Then the tolerance is the minimum projection error 
+                  (as given by empirical_kernel_map::project()) required to cause us to 
+                  include a new vector in the dictionary.  So each time you call add() this 
+                  object basically just computes the projection error for that new sample 
+                  and if it is larger than the tolerance then that new sample becomes part 
+                  of the dictionary.  
         !*/
 
         void clear_dictionary (