Documentation based on 3ffd8c3

_sources/api.md.txt

@@ -88,6 +88,8 @@ You can use the `CombinedLoss` to create arbitrary, weighted sums of losses and
 .. autoclass:: qunfold.GaussianKernelTransformer
 
 .. autoclass:: qunfold.LaplacianKernelTransformer
+
+.. autoclass:: qunfold.GaussianRFFKernelTransformer
 ```
 
 

api.html

@@ -93,6 +93,7 @@
 <li class="toctree-l3"><a class="reference internal" href="#qunfold.EnergyKernelTransformer"><code class="docutils literal notranslate"><span class="pre">EnergyKernelTransformer</span></code></a></li>
 <li class="toctree-l3"><a class="reference internal" href="#qunfold.GaussianKernelTransformer"><code class="docutils literal notranslate"><span class="pre">GaussianKernelTransformer</span></code></a></li>
 <li class="toctree-l3"><a class="reference internal" href="#qunfold.LaplacianKernelTransformer"><code class="docutils literal notranslate"><span class="pre">LaplacianKernelTransformer</span></code></a></li>
+<li class="toctree-l3"><a class="reference internal" href="#qunfold.GaussianRFFKernelTransformer"><code class="docutils literal notranslate"><span class="pre">GaussianRFFKernelTransformer</span></code></a></li>
 </ul>
 </li>
 <li class="toctree-l2"><a class="reference internal" href="#utilities">Utilities</a><ul>
@@ -322,15 +323,16 @@ <h3>Distribution matching<a class="headerlink" href="#distribution-matching" tit
 
 <dl class="py class">
 <dt class="sig sig-object py" id="qunfold.KMM">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">qunfold.</span></span><span class="sig-name descname"><span class="pre">KMM</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">kernel</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'energy'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">sigma</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#qunfold.KMM" title="Permalink to this definition"></a></dt>
-<dd><p>The kernel-based KMM method by Dussap et al. (2023).</p>
-<p>This subclass of <cite>GenericMethod</cite> is instantiated with a <cite>LeastSquaresLoss</cite> and a <cite>KernelTransformer</cite>.</p>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">qunfold.</span></span><span class="sig-name descname"><span class="pre">KMM</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">kernel</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">'energy'</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">sigma</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">n_rff</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1000</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">seed</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="o"><span class="pre">**</span></span><span class="n"><span class="pre">kwargs</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#qunfold.KMM" title="Permalink to this definition"></a></dt>
+<dd><p>The kernel-based KMM method with random Fourier features by Dussap et al. (2023).</p>
+<p>This subclass of <cite>GenericMethod</cite> is instantiated with a <cite>LeastSquaresLoss</cite> and an instance of a <cite>KernelTransformer</cite> sub-class that corresponds to the <cite>kernel</cite> argument.</p>
 <dl class="field-list simple">
 <dt class="field-odd">Parameters<span class="colon">:</span></dt>
 <dd class="field-odd"><ul class="simple">
-<li><p><strong>kernel</strong> (<em>optional</em>) – Which kernel to use. Can be a callable with the signature <cite>(X[y==i], X[y==j]) -&gt; scalar</cite> or one of “energy”, “gaussian”, and “laplacian”. Defaults to “energy”.</p></li>
-<li><p><strong>sigma</strong> (<em>optional</em>) – A smoothing parameter in the <cite>gaussian</cite> and <cite>laplacian</cite> kernels. Defaults to <cite>1</cite>.</p></li>
-<li><p><strong>**kwargs</strong> – Keyword arguments accepted by <cite>GenericMethod</cite>.</p></li>
+<li><p><strong>kernel</strong> (<em>optional</em>) – Which kernel to use. Can be a callable with the signature <cite>(X[y==i], X[y==j]) -&gt; scalar</cite> or one of “energy”, “gaussian”, “laplacian” and “rff”. Defaults to “energy”.</p></li>
+<li><p><strong>sigma</strong> (<em>optional</em>) – A smoothing parameter that is used if <cite>kernel in [“gaussian”, “laplacian”, “rff”]</cite>. Defaults to <cite>1</cite>.</p></li>
+<li><p><strong>n_rff</strong> (<em>optional</em>) – The number of random Fourier features if <cite>kernel == “rff”</cite>. Defaults to <cite>1000</cite>.</p></li>
+<li><p><strong>**kwargs</strong> – Keyword arguments accepted by <cite>GenericMethod</cite>. The <cite>seed</cite> argument also controls the randomness of the random Fourier features if <cite>kernel == “rff”</cite>.</p></li>
 </ul>
 </dd>
 </dl>
@@ -509,14 +511,17 @@ <h2>Feature transformations<a class="headerlink" href="#feature-transformations"
 
 <dl class="py class">
 <dt class="sig sig-object py" id="qunfold.GaussianKernelTransformer">
-<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">qunfold.</span></span><span class="sig-name descname"><span class="pre">GaussianKernelTransformer</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">sigma</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#qunfold.GaussianKernelTransformer" title="Permalink to this definition"></a></dt>
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">qunfold.</span></span><span class="sig-name descname"><span class="pre">GaussianKernelTransformer</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">sigma</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">preprocessor</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#qunfold.GaussianKernelTransformer" title="Permalink to this definition"></a></dt>
 <dd><p>A kernel-based feature transformation, as it is used in <cite>KMM</cite>, that uses the <cite>gaussian</cite> kernel:</p>
 <blockquote>
 <div><p>k(x, y) = exp(-||x - y||^2 / (2σ^2))</p>
 </div></blockquote>
 <dl class="field-list simple">
 <dt class="field-odd">Parameters<span class="colon">:</span></dt>
-<dd class="field-odd"><p><strong>sigma</strong> (<em>optional</em>) – A smoothing parameter of the kernel function. Defaults to <cite>1</cite>.</p>
+<dd class="field-odd"><ul class="simple">
+<li><p><strong>sigma</strong> (<em>optional</em>) – A smoothing parameter of the kernel function. Defaults to <cite>1</cite>.</p></li>
+<li><p><strong>preprocessor</strong> (<em>optional</em>) – Another <cite>AbstractTransformer</cite> that is called before this transformer. Defaults to <cite>None</cite>.</p></li>
+</ul>
 </dd>
 </dl>
 </dd></dl>
@@ -532,6 +537,22 @@ <h2>Feature transformations<a class="headerlink" href="#feature-transformations"
 </dl>
 </dd></dl>
 
+<dl class="py class">
+<dt class="sig sig-object py" id="qunfold.GaussianRFFKernelTransformer">
+<em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">qunfold.</span></span><span class="sig-name descname"><span class="pre">GaussianRFFKernelTransformer</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">sigma</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">n_rff</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">1000</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">preprocessor</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em>, <em class="sig-param"><span class="n"><span class="pre">seed</span></span><span class="o"><span class="pre">=</span></span><span class="default_value"><span class="pre">None</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#qunfold.GaussianRFFKernelTransformer" title="Permalink to this definition"></a></dt>
+<dd><p>An efficient approximation of the <cite>GaussianKernelTransformer</cite>, as it is used in <cite>KMM</cite>, using random Fourier features.</p>
+<dl class="field-list simple">
+<dt class="field-odd">Parameters<span class="colon">:</span></dt>
+<dd class="field-odd"><ul class="simple">
+<li><p><strong>sigma</strong> (<em>optional</em>) – A smoothing parameter of the kernel function. Defaults to <cite>1</cite>.</p></li>
+<li><p><strong>n_rff</strong> (<em>optional</em>) – The number of random Fourier features. Defaults to <cite>1000</cite>.</p></li>
+<li><p><strong>preprocessor</strong> (<em>optional</em>) – Another <cite>AbstractTransformer</cite> that is called before this transformer. Defaults to <cite>None</cite>.</p></li>
+<li><p><strong>seed</strong> (<em>optional</em>) – Controls the randomness of the random Fourier features. Defaults to <cite>None</cite>.</p></li>
+</ul>
+</dd>
+</dl>
+</dd></dl>
+
 </section>
 <section id="utilities">
 <h2>Utilities<a class="headerlink" href="#utilities" title="Permalink to this heading"></a></h2>

developer-guide.html

@@ -257,7 +257,7 @@ <h3>Feature transformations<a class="headerlink" href="#feature-transformations"
 <dl class="py class">
 <dt class="sig sig-object py" id="qunfold.KernelTransformer">
 <em class="property"><span class="pre">class</span><span class="w"> </span></em><span class="sig-prename descclassname"><span class="pre">qunfold.</span></span><span class="sig-name descname"><span class="pre">KernelTransformer</span></span><span class="sig-paren">(</span><em class="sig-param"><span class="n"><span class="pre">kernel</span></span></em><span class="sig-paren">)</span><a class="headerlink" href="#qunfold.KernelTransformer" title="Permalink to this definition"></a></dt>
-<dd><p>A kernel-based feature transformation, as it is used in <cite>KMM</cite>.</p>
+<dd><p>A general kernel-based feature transformation, as it is used in <cite>KMM</cite>. If you intend to use a Gaussian kernel or energy kernel, prefer their dedicated and more efficient implementations over this class.</p>
 <div class="admonition note">
 <p class="admonition-title">Note</p>
 <p>The methods of this transformer do not support setting <cite>average=False</cite>.</p>

genindex.html

@@ -233,6 +233,10 @@ <h2 id="F">F</h2>
         <li><a href="qunfold.html#qunfold.transformers.DistanceTransformer.fit_transform">(qunfold.transformers.DistanceTransformer method)</a>
 </li>
         <li><a href="qunfold.html#qunfold.transformers.EnergyKernelTransformer.fit_transform">(qunfold.transformers.EnergyKernelTransformer method)</a>
+</li>
+        <li><a href="qunfold.html#qunfold.transformers.GaussianKernelTransformer.fit_transform">(qunfold.transformers.GaussianKernelTransformer method)</a>
+</li>
+        <li><a href="qunfold.html#qunfold.transformers.GaussianRFFKernelTransformer.fit_transform">(qunfold.transformers.GaussianRFFKernelTransformer method)</a>
 </li>
         <li><a href="qunfold.html#qunfold.transformers.HistogramTransformer.fit_transform">(qunfold.transformers.HistogramTransformer method)</a>
 </li>
@@ -255,10 +259,16 @@ <h2 id="G">G</h2>
         <li><a href="qunfold.html#qunfold.transformers.GaussianKernelTransformer">(class in qunfold.transformers)</a>
 </li>
       </ul></li>
-      <li><a href="qunfold.tests.html#qunfold.tests.generate_data">generate_data() (in module qunfold.tests)</a>
+      <li><a href="api.html#qunfold.GaussianRFFKernelTransformer">GaussianRFFKernelTransformer (class in qunfold)</a>
+
+      <ul>
+        <li><a href="qunfold.html#qunfold.transformers.GaussianRFFKernelTransformer">(class in qunfold.transformers)</a>
 </li>
+      </ul></li>
   </ul></td>
   <td style="width: 33%; vertical-align: top;"><ul>
+      <li><a href="qunfold.tests.html#qunfold.tests.generate_data">generate_data() (in module qunfold.tests)</a>
+</li>
       <li><a href="api.html#qunfold.GenericMethod">GenericMethod (class in qunfold)</a>
 
       <ul>
@@ -321,12 +331,8 @@ <h2 id="J">J</h2>
 <h2 id="K">K</h2>
 <table style="width: 100%" class="indextable genindextable"><tr>
   <td style="width: 33%; vertical-align: top;"><ul>
-      <li><a href="qunfold.html#qunfold.transformers.GaussianKernelTransformer.kernel">kernel (qunfold.transformers.GaussianKernelTransformer property)</a>
-
-      <ul>
-        <li><a href="qunfold.html#qunfold.transformers.LaplacianKernelTransformer.kernel">(qunfold.transformers.LaplacianKernelTransformer property)</a>
+      <li><a href="qunfold.html#qunfold.transformers.LaplacianKernelTransformer.kernel">kernel (qunfold.transformers.LaplacianKernelTransformer property)</a>
 </li>
-      </ul></li>
       <li><a href="developer-guide.html#qunfold.KernelTransformer">KernelTransformer (class in qunfold)</a>
 
       <ul>
@@ -647,10 +653,10 @@ <h2 id="T">T</h2>
 </li>
       <li><a href="qunfold.tests.html#qunfold.tests.TestMethods">TestMethods (class in qunfold.tests)</a>
 </li>
-  </ul></td>
-  <td style="width: 33%; vertical-align: top;"><ul>
       <li><a href="qunfold.tests.html#qunfold.tests.TestQuaPyWrapper">TestQuaPyWrapper (class in qunfold.tests)</a>
 </li>
+  </ul></td>
+  <td style="width: 33%; vertical-align: top;"><ul>
       <li><a href="api.html#qunfold.TikhonovRegularization">TikhonovRegularization (class in qunfold)</a>
 
       <ul>
@@ -671,6 +677,10 @@ <h2 id="T">T</h2>
         <li><a href="qunfold.html#qunfold.transformers.DistanceTransformer.transform">(qunfold.transformers.DistanceTransformer method)</a>
 </li>
         <li><a href="qunfold.html#qunfold.transformers.EnergyKernelTransformer.transform">(qunfold.transformers.EnergyKernelTransformer method)</a>
+</li>
+        <li><a href="qunfold.html#qunfold.transformers.GaussianKernelTransformer.transform">(qunfold.transformers.GaussianKernelTransformer method)</a>
+</li>
+        <li><a href="qunfold.html#qunfold.transformers.GaussianRFFKernelTransformer.transform">(qunfold.transformers.GaussianRFFKernelTransformer method)</a>
 </li>
         <li><a href="qunfold.html#qunfold.transformers.HistogramTransformer.transform">(qunfold.transformers.HistogramTransformer method)</a>
 </li>

modules.html

@@ -173,7 +173,13 @@ <h1>qunfold<a class="headerlink" href="#qunfold" title="Permalink to this headin
 </ul>
 </li>
 <li class="toctree-l3"><a class="reference internal" href="qunfold.html#qunfold.transformers.GaussianKernelTransformer"><code class="docutils literal notranslate"><span class="pre">GaussianKernelTransformer</span></code></a><ul>
-<li class="toctree-l4"><a class="reference internal" href="qunfold.html#qunfold.transformers.GaussianKernelTransformer.kernel"><code class="docutils literal notranslate"><span class="pre">GaussianKernelTransformer.kernel</span></code></a></li>
+<li class="toctree-l4"><a class="reference internal" href="qunfold.html#qunfold.transformers.GaussianKernelTransformer.fit_transform"><code class="docutils literal notranslate"><span class="pre">GaussianKernelTransformer.fit_transform()</span></code></a></li>
+<li class="toctree-l4"><a class="reference internal" href="qunfold.html#qunfold.transformers.GaussianKernelTransformer.transform"><code class="docutils literal notranslate"><span class="pre">GaussianKernelTransformer.transform()</span></code></a></li>
+</ul>
+</li>
+<li class="toctree-l3"><a class="reference internal" href="qunfold.html#qunfold.transformers.GaussianRFFKernelTransformer"><code class="docutils literal notranslate"><span class="pre">GaussianRFFKernelTransformer</span></code></a><ul>
+<li class="toctree-l4"><a class="reference internal" href="qunfold.html#qunfold.transformers.GaussianRFFKernelTransformer.fit_transform"><code class="docutils literal notranslate"><span class="pre">GaussianRFFKernelTransformer.fit_transform()</span></code></a></li>
+<li class="toctree-l4"><a class="reference internal" href="qunfold.html#qunfold.transformers.GaussianRFFKernelTransformer.transform"><code class="docutils literal notranslate"><span class="pre">GaussianRFFKernelTransformer.transform()</span></code></a></li>
 </ul>
 </li>
 <li class="toctree-l3"><a class="reference internal" href="qunfold.html#qunfold.transformers.HistogramTransformer"><code class="docutils literal notranslate"><span class="pre">HistogramTransformer</span></code></a><ul>