API updated

doc/build/html/.buildinfo

@@ -1,4 +1,4 @@
 # Sphinx build info version 1
 # This file hashes the configuration used when building these files. When it is not found, a full rebuild will be done.
-config: 2abc3c3e7678fef7f1895889019e9c34
+config: 671334558ad756f088d7c8b8718316be
 tags: 645f666f9bcd5a90fca523b33c5a78b7

doc/build/html/CV.html

@@ -170,7 +170,7 @@ <h3>Navigation</h3>
     </div>
     <div class="footer">
         &copy; Copyright 2013, Marion Neumann, Shan Huang, Daniel Marthaler, Kristian Kersting.
-      Created using <a href="http://sphinx-doc.org/">Sphinx</a> 1.2.
+      Created using <a href="http://sphinx-doc.org/">Sphinx</a> 1.2.2.
     </div>
   </body>
 </html>

doc/build/html/Examples.html

@@ -54,9 +54,9 @@ <h3>Navigation</h3>
 
   <div class="section" id="demos">
 <h1>Demos<a class="headerlink" href="#demos" title="Permalink to this headline">¶</a></h1>
-<p>There are several demos exemplifying the use of pyGPs for various Gaussian process (<img class="math" src="_images/math/192d69ed9c7bd67b3f34486786531bae08ba2d88.png" alt="GP"/>) tasks.
-We recommend to first go through <em>Basic GP Regression</em> which introduces the <img class="math" src="_images/math/192d69ed9c7bd67b3f34486786531bae08ba2d88.png" alt="GP"/> regression model.
-Basic regression is the most intuitive and simplest learning task feasable with <img class="math" src="_images/math/502c8d8d8b92a24e5186c2d0d79f44b1d97e373b.png" alt="GPs"/>.
+<p>There are several demos exemplifying the use of pyGPs for various Gaussian process (<img class="math" src="_images/math/bc008ec23119a8f24d723e0616fee9a6f9a87cd2.png" alt="GP"/>) tasks.
+We recommend to first go through <em>Basic GP Regression</em> which introduces the <img class="math" src="_images/math/bc008ec23119a8f24d723e0616fee9a6f9a87cd2.png" alt="GP"/> regression model.
+Basic regression is the most intuitive and simplest learning task feasable with <img class="math" src="_images/math/296cca5da127cd61508d98f42eb470f637e9912c.png" alt="GPs"/>.
 The other demos will then provide a general insight into more advanced functionalities of the package.
 You will also find the implementation of the demos in the <a class="reference external" href="https://github.com/marionmari/pyGPs">source</a> folder under <a class="reference external" href="https://github.com/marionmari/pyGPs/tree/master/pyGPs/Demo">pyGPs/Demo</a>.</p>
 <p>The Demos give some theoretical explanations. Further, it is useful to have a look at our documentation on <a class="reference external" href="Kernels.html">Kernels &amp; Means</a> and <a class="reference external" href="Opts.html">Optimizers</a>.</p>
@@ -141,7 +141,7 @@ <h3>Navigation</h3>
     </div>
     <div class="footer">
         &copy; Copyright 2013, Marion Neumann, Shan Huang, Daniel Marthaler, Kristian Kersting.
-      Created using <a href="http://sphinx-doc.org/">Sphinx</a> 1.2.
+      Created using <a href="http://sphinx-doc.org/">Sphinx</a> 1.2.2.
     </div>
   </body>
 </html>

doc/build/html/GPC.html

@@ -67,17 +67,17 @@ <h2>Load data<a class="headerlink" href="#load-data" title="Permalink to this he
 <span class="n">z</span> <span class="o">=</span> <span class="n">demoData</span><span class="p">[</span><span class="s">&#39;xstar&#39;</span><span class="p">]</span>        <span class="c"># test data</span>
 </pre></div>
 </div>
-<p>The <img class="math" src="_images/math/aaa4de95cc6371d125d5cf773cdd9f9913dd1d8e.png" alt="120"/> data points were generated from two Gaussians with different means and covariances. One Gaussian is isotropic and contains
-<img class="math" src="_images/math/38bd36e353df831303c0de895a9ec614cf3e7202.png" alt="2/3"/> of the data (blue), the other is highly correlated and contains <img class="math" src="_images/math/217aedbdc339bacc8ba075a2ec16902b098194e3.png" alt="1/3"/> of the points (red).
-Note, that the labels for the targets are specified to be <img class="math" src="_images/math/7cfbd35086ffefb83637d42166582b43c8a1ff4a.png" alt="\pm 1"/> (and not <img class="math" src="_images/math/e15d84dfcdccf6b0d8fb485020852b5b0f4ea097.png" alt="0/1"/>).</p>
+<p>The <img class="math" src="_images/math/1c1fc25eb56b788647384456e0e1a8dade41a0c1.png" alt="120"/> data points were generated from two Gaussians with different means and covariances. One Gaussian is isotropic and contains
+<img class="math" src="_images/math/62cfeebb5ed187b89fa9e16c9413edbb3509ae74.png" alt="2/3"/> of the data (blue), the other is highly correlated and contains <img class="math" src="_images/math/6ce2356fc9366d51aa33bf8c380f04bdda621a67.png" alt="1/3"/> of the points (red).
+Note, that the labels for the targets are specified to be <img class="math" src="_images/math/5afef8883a1e1142173e76df8d16e764b2779782.png" alt="\pm 1"/> (and not <img class="math" src="_images/math/26b1a580c36b3c8ff2d4b4419ad8a7d054213501.png" alt="0/1"/>).</p>
 <p>In the plot, we superimpose the data points with the posterior equi-probability contour lines for the probability of the second class
 given complete information about the generating mechanism.</p>
 <div class="figure align-center">
 <a class="reference internal image-reference" href="_images/d2_1.png"><img alt="_images/d2_1.png" src="_images/d2_1.png" style="width: 560.0px; height: 420.0px;" /></a>
 </div>
 </div>
 <div class="section" id="first-example-state-default-values">
-<h2>First example <img class="math" src="_images/math/e12b6767375342ed57d27678e3ea1cdb97f47e15.png" alt="\rightarrow"/> state default values<a class="headerlink" href="#first-example-state-default-values" title="Permalink to this headline">¶</a></h2>
+<h2>First example <img class="math" src="_images/math/a9c4c6156d25f42923975ce449aadad9848ed7dc.png" alt="\rightarrow"/> state default values<a class="headerlink" href="#first-example-state-default-values" title="Permalink to this headline">¶</a></h2>
 <p>Again, lets see the simplest use of gp classification at first</p>
 <div class="highlight-python"><div class="highlight"><pre><span class="n">model</span> <span class="o">=</span> <span class="n">pyGPs</span><span class="o">.</span><span class="n">gp</span><span class="o">.</span><span class="n">GPC</span><span class="p">()</span>             <span class="c"># binary classification (default inference method: EP)</span>
 <span class="n">model</span><span class="o">.</span><span class="n">fit</span><span class="p">(</span><span class="n">x</span><span class="p">,</span> <span class="n">y</span><span class="p">)</span>              <span class="c"># fit default model (mean zero &amp; rbf kernel) with data</span>
@@ -88,8 +88,8 @@ <h2>First example <img class="math" src="_images/math/e12b6767375342ed57d27678e3
 <p>Note, that inference is done via expectation propagation (EP) approximation by deault. How to set inference to Laplace approximation, see <a class="reference internal" href="#more-on-gpc"><em>A bit more things you can do</em></a>.</p>
 </div>
 <div class="section" id="second-example-gp-classification">
-<h2>Second example <img class="math" src="_images/math/e12b6767375342ed57d27678e3ea1cdb97f47e15.png" alt="\rightarrow"/> GP classification<a class="headerlink" href="#second-example-gp-classification" title="Permalink to this headline">¶</a></h2>
-<p>So we first state the model to be <img class="math" src="_images/math/192d69ed9c7bd67b3f34486786531bae08ba2d88.png" alt="GP"/> classification now:</p>
+<h2>Second example <img class="math" src="_images/math/a9c4c6156d25f42923975ce449aadad9848ed7dc.png" alt="\rightarrow"/> GP classification<a class="headerlink" href="#second-example-gp-classification" title="Permalink to this headline">¶</a></h2>
+<p>So we first state the model to be <img class="math" src="_images/math/bc008ec23119a8f24d723e0616fee9a6f9a87cd2.png" alt="GP"/> classification now:</p>
 <div class="highlight-python"><div class="highlight"><pre><span class="n">model</span> <span class="o">=</span> <span class="n">pyGPs</span><span class="o">.</span><span class="n">GPC</span><span class="p">()</span>
 </pre></div>
 </div>
@@ -107,8 +107,8 @@ <h2>Second example <img class="math" src="_images/math/e12b6767375342ed57d27678e
 </div>
 <p><strong>[Theory]</strong>
 In this example, we used an RBF kernel (squared exponential covariance function) with automatic relevance determination (ARD). This covariance function has one
-characteristic length-scale parameter for each dimension of the input space (here <img class="math" src="_images/math/41c544263a265ff15498ee45f7392c5f86c6d151.png" alt="2"/> in total), and a signal magnitude parameter, resulting in
-a total of <img class="math" src="_images/math/7cde695f2e4542fd01f860a89189f47a27143b66.png" alt="3"/> hyperparameters. ARD with separate length-scales for each input dimension is a very powerful tool to learn which
+characteristic length-scale parameter for each dimension of the input space (here <img class="math" src="_images/math/15c663954a3e059d1f876bc8a4621de376038c96.png" alt="2"/> in total), and a signal magnitude parameter, resulting in
+a total of <img class="math" src="_images/math/b9b358d9bbdf54c3d9aef7554638822d996c21ea.png" alt="3"/> hyperparameters. ARD with separate length-scales for each input dimension is a very powerful tool to learn which
 inputs are important for the predictions: if length-scales are short, input dimensions are very important, and when they grow very large
 (compared to the spread of the data), the corresponding input dimensions will be mostly ignored.</p>
 <p>Note, <em>pyGPs.GPC().plot()</em> is a toy method for 2-d data:</p>
@@ -117,7 +117,7 @@ <h2>Second example <img class="math" src="_images/math/e12b6767375342ed57d27678e
 </div>
 <p>The contour plot for the predictive distribution is shown below. Note, that the predictive
 probability is fairly close to the probabilities of the generating process in regions of high data density. Note also, that as you move
-away from the data, the probability approaches <img class="math" src="_images/math/217aedbdc339bacc8ba075a2ec16902b098194e3.png" alt="1/3"/>, the overall class probability.</p>
+away from the data, the probability approaches <img class="math" src="_images/math/6ce2356fc9366d51aa33bf8c380f04bdda621a67.png" alt="1/3"/>, the overall class probability.</p>
 <div class="figure align-center">
 <a class="reference internal image-reference" href="_images/d2_2.png"><img alt="_images/d2_2.png" src="_images/d2_2.png" style="width: 560.0px; height: 420.0px;" /></a>
 </div>
@@ -143,8 +143,8 @@ <h3><a href="index.html">Table Of Contents</a></h3>
   <ul>
 <li><a class="reference internal" href="#">Basic Classification</a><ul>
 <li><a class="reference internal" href="#load-data">Load data</a></li>
-<li><a class="reference internal" href="#first-example-state-default-values">First example <img class="math" src="_images/math/e12b6767375342ed57d27678e3ea1cdb97f47e15.png" alt="\rightarrow"/> state default values</a></li>
-<li><a class="reference internal" href="#second-example-gp-classification">Second example <img class="math" src="_images/math/e12b6767375342ed57d27678e3ea1cdb97f47e15.png" alt="\rightarrow"/> GP classification</a></li>
+<li><a class="reference internal" href="#first-example-state-default-values">First example <img class="math" src="_images/math/a9c4c6156d25f42923975ce449aadad9848ed7dc.png" alt="\rightarrow"/> state default values</a></li>
+<li><a class="reference internal" href="#second-example-gp-classification">Second example <img class="math" src="_images/math/a9c4c6156d25f42923975ce449aadad9848ed7dc.png" alt="\rightarrow"/> GP classification</a></li>
 <li><a class="reference internal" href="#a-bit-more-things-you-can-do">A bit more things you can do</a></li>
 </ul>
 </li>
@@ -199,7 +199,7 @@ <h3>Navigation</h3>
     </div>
     <div class="footer">
         &copy; Copyright 2013, Marion Neumann, Shan Huang, Daniel Marthaler, Kristian Kersting.
-      Created using <a href="http://sphinx-doc.org/">Sphinx</a> 1.2.
+      Created using <a href="http://sphinx-doc.org/">Sphinx</a> 1.2.2.
     </div>
   </body>
 </html>