Commit 7679e6aa authored by Davis King's avatar Davis King

updated the docs

--HG--
extra : convert_revision : svn%3Afdd8eb12-d10e-0410-9acb-85c331704f74/trunk%403245
parent a36189a5
...@@ -126,8 +126,8 @@ Davis E. King. <a href="http://www.jmlr.org/papers/volume10/king09a/king09a.pdf" ...@@ -126,8 +126,8 @@ Davis E. King. <a href="http://www.jmlr.org/papers/volume10/king09a/king09a.pdf"
<item>cross_validate_trainer</item> <item>cross_validate_trainer</item>
<item>cross_validate_trainer_threaded</item> <item>cross_validate_trainer_threaded</item>
<item>pick_initial_centers</item> <item>pick_initial_centers</item>
<item>rank_features_rbf</item> <item>find_gamma_with_big_centroid_gap</item>
<item>verbose_rank_features_rbf</item> <item>verbose_find_gamma_with_big_centroid_gap</item>
</section> </section>
</top> </top>
...@@ -747,14 +747,6 @@ Davis E. King. <a href="http://www.jmlr.org/papers/volume10/king09a/king09a.pdf" ...@@ -747,14 +747,6 @@ Davis E. King. <a href="http://www.jmlr.org/papers/volume10/king09a/king09a.pdf"
does this by computing the distance between the centroids of both classes in kernel defined does this by computing the distance between the centroids of both classes in kernel defined
feature space. Good features are then ones that result in the biggest separation between feature space. Good features are then ones that result in the biggest separation between
the two centroids. the two centroids.
<p>
Note that there are convenience wrappers for calling this function.
They are the <a href="#rank_features_rbf">rank_features_rbf</a> and
<a href="#verbose_rank_features_rbf">verbose_rank_features_rbf</a> functions.
They automatically create the <a href="#kcentroid">kcentroid</a> required to
use this function and fill out the relevant parameters for you with a reasonable set
of defaults.
</p>
</description> </description>
<examples> <examples>
<example>rank_features_ex.cpp.html</example> <example>rank_features_ex.cpp.html</example>
...@@ -765,13 +757,14 @@ Davis E. King. <a href="http://www.jmlr.org/papers/volume10/king09a/king09a.pdf" ...@@ -765,13 +757,14 @@ Davis E. King. <a href="http://www.jmlr.org/papers/volume10/king09a/king09a.pdf"
<!-- ************************************************************************* --> <!-- ************************************************************************* -->
<component> <component>
<name>rank_features_rbf</name> <name>find_gamma_with_big_centroid_gap</name>
<file>dlib/svm.h</file> <file>dlib/svm.h</file>
<spec_file link="true">dlib/svm/feature_ranking_abstract.h</spec_file> <spec_file link="true">dlib/svm/feature_ranking_abstract.h</spec_file>
<description> <description>
This is just a wrapper for the <a href="#rank_features">rank_features</a> function. It This is a function that tries to pick a reasonable default value for the
sets up the ranking to use the <a href="#radial_basis_kernel">radial_basis_kernel</a> and gamma parameter of the <a href="#radial_basis_kernel">radial_basis_kernel</a>. It
automatically picks the gamma parameter of this kernel. picks the parameter that gives the largest separation between the centroids, in
kernel feature space, of two classes of data.
</description> </description>
<examples> <examples>
<example>rank_features_ex.cpp.html</example> <example>rank_features_ex.cpp.html</example>
...@@ -782,11 +775,12 @@ Davis E. King. <a href="http://www.jmlr.org/papers/volume10/king09a/king09a.pdf" ...@@ -782,11 +775,12 @@ Davis E. King. <a href="http://www.jmlr.org/papers/volume10/king09a/king09a.pdf"
<!-- ************************************************************************* --> <!-- ************************************************************************* -->
<component> <component>
<name>verbose_rank_features_rbf</name> <name>verbose_find_gamma_with_big_centroid_gap</name>
<file>dlib/svm.h</file> <file>dlib/svm.h</file>
<spec_file link="true">dlib/svm/feature_ranking_abstract.h</spec_file> <spec_file link="true">dlib/svm/feature_ranking_abstract.h</spec_file>
<description> <description>
This is just a version of <a href="#rank_features_rbf">rank_features_rbf</a> that is This is just a version of
<a href="#find_gamma_with_big_centroid_gap">find_gamma_with_big_centroid_gap</a> that is
verbose in the sense that it prints out status messages as it proceeds. verbose in the sense that it prints out status messages as it proceeds.
</description> </description>
<examples> <examples>
......
...@@ -28,6 +28,8 @@ ...@@ -28,6 +28,8 @@
<term file="dlib/optimization/optimization_abstract.h.html" name="central_differences"/> <term file="dlib/optimization/optimization_abstract.h.html" name="central_differences"/>
<term file="dlib/optimization/optimization_abstract.h.html" name="line_search_funct"/> <term file="dlib/optimization/optimization_abstract.h.html" name="line_search_funct"/>
<term file="dlib/optimization/optimization_line_search_abstract.h.html" name="optimize_single_variable_failure"/>
<term file="dlib/optimization/optimization_bobyqa_abstract.h.html" name="bobyqa_failure"/>
<term file="algorithms.html" name="derivative"/> <term file="algorithms.html" name="derivative"/>
<term file="algorithms.html" name="make_line_search_function"/> <term file="algorithms.html" name="make_line_search_function"/>
...@@ -138,8 +140,8 @@ ...@@ -138,8 +140,8 @@
<term file="ml.html" name="kkmeans"/> <term file="ml.html" name="kkmeans"/>
<term file="ml.html" name="pick_initial_centers"/> <term file="ml.html" name="pick_initial_centers"/>
<term file="ml.html" name="rank_features"/> <term file="ml.html" name="rank_features"/>
<term file="ml.html" name="rank_features_rbf"/> <term file="ml.html" name="find_gamma_with_big_centroid_gap"/>
<term file="ml.html" name="verbose_rank_features_rbf"/> <term file="ml.html" name="verbose_find_gamma_with_big_centroid_gap"/>
<term file="dlib/svm/svm_abstract.h.html" name="maximum_nu"/> <term file="dlib/svm/svm_abstract.h.html" name="maximum_nu"/>
<term file="ml.html" name="null_trainer"/> <term file="ml.html" name="null_trainer"/>
......
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment