moreo
/
QuaPy
1
0
Fork 1
Code Issues Pull Requests Releases Wiki Activity
QuaPy/docs/build/html/_modules/quapy/error.html

664 lines
67 KiB
HTML
Raw Blame History

<!DOCTYPE html>
<html class="writer-html5" lang="en" data-content_root="../../">
<head>
<meta charset="utf-8" />
<meta name="viewport" content="width=device-width, initial-scale=1.0" />
<title>quapy.error &mdash; QuaPy: A Python-based open-source framework for quantification 0.2.1 documentation</title>
<link rel="stylesheet" type="text/css" href="../../_static/pygments.css?v=b86133f3" />
<link rel="stylesheet" type="text/css" href="../../_static/css/theme.css?v=9edc463e" />
<script src="../../_static/jquery.js?v=5d32c60e"></script>
<script src="../../_static/_sphinx_javascript_frameworks_compat.js?v=2cd50e6c"></script>
<script src="../../_static/documentation_options.js?v=37f418d5"></script>
<script src="../../_static/doctools.js?v=fd6eb6e6"></script>
<script src="../../_static/sphinx_highlight.js?v=6ffebe34"></script>
<script src="../../_static/js/theme.js"></script>
<link rel="index" title="Index" href="../../genindex.html" />
<link rel="search" title="Search" href="../../search.html" />
</head>
<body class="wy-body-for-nav">
<div class="wy-grid-for-nav">
<nav data-toggle="wy-nav-shift" class="wy-nav-side">
<div class="wy-side-scroll">
<div class="wy-side-nav-search" >
<a href="../../index.html" class="icon icon-home">
QuaPy: A Python-based open-source framework for quantification
</a>
<div role="search">
<form id="rtd-search-form" class="wy-form" action="../../search.html" method="get">
<input type="text" name="q" placeholder="Search docs" aria-label="Search docs" />
<input type="hidden" name="check_keywords" value="yes" />
<input type="hidden" name="area" value="default" />
</form>
</div>
</div><div class="wy-menu wy-menu-vertical" data-spy="affix" role="navigation" aria-label="Navigation menu">
<ul>
<li class="toctree-l1"><a class="reference internal" href="../../index.html">Quickstart</a></li>
</ul>
<ul>
<li class="toctree-l1"><a class="reference internal" href="../../manuals.html">Manuals</a></li>
</ul>
<ul>
<li class="toctree-l1"><a class="reference internal" href="../../quapy.html">API</a></li>
</ul>
</div>
</div>
</nav>
<section data-toggle="wy-nav-shift" class="wy-nav-content-wrap"><nav class="wy-nav-top" aria-label="Mobile navigation menu" >
<i data-toggle="wy-nav-top" class="fa fa-bars"></i>
<a href="../../index.html">QuaPy: A Python-based open-source framework for quantification</a>
</nav>
<div class="wy-nav-content">
<div class="rst-content">
<div role="navigation" aria-label="Page navigation">
<ul class="wy-breadcrumbs">
<li><a href="../../index.html" class="icon icon-home" aria-label="Home"></a></li>
<li class="breadcrumb-item"><a href="../index.html">Module code</a></li>
<li class="breadcrumb-item active">quapy.error</li>
<li class="wy-breadcrumbs-aside">
</li>
</ul>
<hr/>
</div>
<div role="main" class="document" itemscope="itemscope" itemtype="http://schema.org/Article">
<div itemprop="articleBody">
<h1>Source code for quapy.error</h1><div class="highlight"><pre>
<span></span><span class="sd">&quot;&quot;&quot;Implementation of error measures used for quantification&quot;&quot;&quot;</span>
<span class="kn">import</span><span class="w"> </span><span class="nn">numpy</span><span class="w"> </span><span class="k">as</span><span class="w"> </span><span class="nn">np</span>
<span class="kn">from</span><span class="w"> </span><span class="nn">sklearn.metrics</span><span class="w"> </span><span class="kn">import</span> <span class="n">f1_score</span>
<span class="kn">import</span><span class="w"> </span><span class="nn">quapy</span><span class="w"> </span><span class="k">as</span><span class="w"> </span><span class="nn">qp</span>
<div class="viewcode-block" id="from_name">
<a class="viewcode-back" href="../../quapy.html#quapy.error.from_name">[docs]</a>
<span class="k">def</span><span class="w"> </span><span class="nf">from_name</span><span class="p">(</span><span class="n">err_name</span><span class="p">):</span>
<span class="w"> </span><span class="sd">&quot;&quot;&quot;Gets an error function from its name. E.g., `from_name(&quot;mae&quot;)`</span>
<span class="sd"> will return function :meth:`quapy.error.mae`</span>
<span class="sd"> :param err_name: string, the error name</span>
<span class="sd"> :return: a callable implementing the requested error</span>
<span class="sd"> &quot;&quot;&quot;</span>
<span class="k">assert</span> <span class="n">err_name</span> <span class="ow">in</span> <span class="n">ERROR_NAMES</span><span class="p">,</span> <span class="sa">f</span><span class="s1">&#39;unknown error </span><span class="si">{</span><span class="n">err_name</span><span class="si">}</span><span class="s1">&#39;</span>
<span class="n">callable_error</span> <span class="o">=</span> <span class="nb">globals</span><span class="p">()[</span><span class="n">err_name</span><span class="p">]</span>
<span class="k">return</span> <span class="n">callable_error</span></div>
<div class="viewcode-block" id="f1e">
<a class="viewcode-back" href="../../quapy.html#quapy.error.f1e">[docs]</a>
<span class="k">def</span><span class="w"> </span><span class="nf">f1e</span><span class="p">(</span><span class="n">y_true</span><span class="p">,</span> <span class="n">y_pred</span><span class="p">):</span>
<span class="w"> </span><span class="sd">&quot;&quot;&quot;F1 error: simply computes the error in terms of macro :math:`F_1`, i.e.,</span>
<span class="sd"> :math:`1-F_1^M`, where :math:`F_1` is the harmonic mean of precision and recall,</span>
<span class="sd"> defined as :math:`\\frac{2tp}{2tp+fp+fn}`, with `tp`, `fp`, and `fn` standing</span>
<span class="sd"> for true positives, false positives, and false negatives, respectively.</span>
<span class="sd"> `Macro` averaging means the :math:`F_1` is computed for each category independently,</span>
<span class="sd"> and then averaged.</span>
<span class="sd"> :param y_true: array-like of true labels</span>
<span class="sd"> :param y_pred: array-like of predicted labels</span>
<span class="sd"> :return: :math:`1-F_1^M`</span>
<span class="sd"> &quot;&quot;&quot;</span>
<span class="k">return</span> <span class="mf">1.</span> <span class="o">-</span> <span class="n">f1_score</span><span class="p">(</span><span class="n">y_true</span><span class="p">,</span> <span class="n">y_pred</span><span class="p">,</span> <span class="n">average</span><span class="o">=</span><span class="s1">&#39;macro&#39;</span><span class="p">)</span></div>
<div class="viewcode-block" id="acce">
<a class="viewcode-back" href="../../quapy.html#quapy.error.acce">[docs]</a>
<span class="k">def</span><span class="w"> </span><span class="nf">acce</span><span class="p">(</span><span class="n">y_true</span><span class="p">,</span> <span class="n">y_pred</span><span class="p">):</span>
<span class="w"> </span><span class="sd">&quot;&quot;&quot;Computes the error in terms of 1-accuracy. The accuracy is computed as</span>
<span class="sd"> :math:`\\frac{tp+tn}{tp+fp+fn+tn}`, with `tp`, `fp`, `fn`, and `tn` standing</span>
<span class="sd"> for true positives, false positives, false negatives, and true negatives,</span>
<span class="sd"> respectively</span>
<span class="sd"> :param y_true: array-like of true labels</span>
<span class="sd"> :param y_pred: array-like of predicted labels</span>
<span class="sd"> :return: 1-accuracy</span>
<span class="sd"> &quot;&quot;&quot;</span>
<span class="k">return</span> <span class="mf">1.</span> <span class="o">-</span> <span class="p">(</span><span class="n">y_true</span> <span class="o">==</span> <span class="n">y_pred</span><span class="p">)</span><span class="o">.</span><span class="n">mean</span><span class="p">()</span></div>
<div class="viewcode-block" id="mae">
<a class="viewcode-back" href="../../quapy.html#quapy.error.mae">[docs]</a>
<span class="k">def</span><span class="w"> </span><span class="nf">mae</span><span class="p">(</span><span class="n">prevs_true</span><span class="p">,</span> <span class="n">prevs_hat</span><span class="p">):</span>
<span class="w"> </span><span class="sd">&quot;&quot;&quot;Computes the mean absolute error (see :meth:`quapy.error.ae`) across the sample pairs.</span>
<span class="sd"> :param prevs_true: array-like of shape `(n_samples, n_classes,)` with the true prevalence values</span>
<span class="sd"> :param prevs_hat: array-like of shape `(n_samples, n_classes,)` with the predicted</span>
<span class="sd"> prevalence values</span>
<span class="sd"> :return: mean absolute error</span>
<span class="sd"> &quot;&quot;&quot;</span>
<span class="k">return</span> <span class="n">ae</span><span class="p">(</span><span class="n">prevs_true</span><span class="p">,</span> <span class="n">prevs_hat</span><span class="p">)</span><span class="o">.</span><span class="n">mean</span><span class="p">()</span></div>
<div class="viewcode-block" id="ae">
<a class="viewcode-back" href="../../quapy.html#quapy.error.ae">[docs]</a>
<span class="k">def</span><span class="w"> </span><span class="nf">ae</span><span class="p">(</span><span class="n">prevs_true</span><span class="p">,</span> <span class="n">prevs_hat</span><span class="p">):</span>
<span class="w"> </span><span class="sd">&quot;&quot;&quot;Computes the absolute error between the two prevalence vectors.</span>
<span class="sd"> Absolute error between two prevalence vectors :math:`p` and :math:`\\hat{p}` is computed as</span>
<span class="sd"> :math:`AE(p,\\hat{p})=\\frac{1}{|\\mathcal{Y}|}\\sum_{y\\in \\mathcal{Y}}|\\hat{p}(y)-p(y)|`,</span>
<span class="sd"> where :math:`\\mathcal{Y}` are the classes of interest.</span>
<span class="sd"> :param prevs_true: array-like of shape `(n_classes,)` with the true prevalence values</span>
<span class="sd"> :param prevs_hat: array-like of shape `(n_classes,)` with the predicted prevalence values</span>
<span class="sd"> :return: absolute error</span>
<span class="sd"> &quot;&quot;&quot;</span>
<span class="n">prevs_true</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">asarray</span><span class="p">(</span><span class="n">prevs_true</span><span class="p">)</span>
<span class="n">prevs_hat</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">asarray</span><span class="p">(</span><span class="n">prevs_hat</span><span class="p">)</span>
<span class="k">assert</span> <span class="n">prevs_true</span><span class="o">.</span><span class="n">shape</span> <span class="o">==</span> <span class="n">prevs_hat</span><span class="o">.</span><span class="n">shape</span><span class="p">,</span> <span class="sa">f</span><span class="s1">&#39;wrong shape </span><span class="si">{</span><span class="n">prevs_true</span><span class="o">.</span><span class="n">shape</span><span class="si">}</span><span class="s1"> vs. </span><span class="si">{</span><span class="n">prevs_hat</span><span class="o">.</span><span class="n">shape</span><span class="si">}</span><span class="s1">&#39;</span>
<span class="k">return</span> <span class="nb">abs</span><span class="p">(</span><span class="n">prevs_hat</span> <span class="o">-</span> <span class="n">prevs_true</span><span class="p">)</span><span class="o">.</span><span class="n">mean</span><span class="p">(</span><span class="n">axis</span><span class="o">=-</span><span class="mi">1</span><span class="p">)</span></div>
<div class="viewcode-block" id="nae">
<a class="viewcode-back" href="../../quapy.html#quapy.error.nae">[docs]</a>
<span class="k">def</span><span class="w"> </span><span class="nf">nae</span><span class="p">(</span><span class="n">prevs_true</span><span class="p">,</span> <span class="n">prevs_hat</span><span class="p">):</span>
<span class="w"> </span><span class="sd">&quot;&quot;&quot;Computes the normalized absolute error between the two prevalence vectors.</span>
<span class="sd"> Normalized absolute error between two prevalence vectors :math:`p` and :math:`\\hat{p}` is computed as</span>
<span class="sd"> :math:`NAE(p,\\hat{p})=\\frac{AE(p,\\hat{p})}{z_{AE}}`,</span>
<span class="sd"> where :math:`z_{AE}=\\frac{2(1-\\min_{y\\in \\mathcal{Y}} p(y))}{|\\mathcal{Y}|}`, and :math:`\\mathcal{Y}`</span>
<span class="sd"> are the classes of interest.</span>
<span class="sd"> :param prevs_true: array-like of shape `(n_classes,)` with the true prevalence values</span>
<span class="sd"> :param prevs_hat: array-like of shape `(n_classes,)` with the predicted prevalence values</span>
<span class="sd"> :return: normalized absolute error</span>
<span class="sd"> &quot;&quot;&quot;</span>
<span class="n">prevs_true</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">asarray</span><span class="p">(</span><span class="n">prevs_true</span><span class="p">)</span>
<span class="n">prevs_hat</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">asarray</span><span class="p">(</span><span class="n">prevs_hat</span><span class="p">)</span>
<span class="k">assert</span> <span class="n">prevs_true</span><span class="o">.</span><span class="n">shape</span> <span class="o">==</span> <span class="n">prevs_hat</span><span class="o">.</span><span class="n">shape</span><span class="p">,</span> <span class="sa">f</span><span class="s1">&#39;wrong shape </span><span class="si">{</span><span class="n">prevs_true</span><span class="o">.</span><span class="n">shape</span><span class="si">}</span><span class="s1"> vs. </span><span class="si">{</span><span class="n">prevs_hat</span><span class="o">.</span><span class="n">shape</span><span class="si">}</span><span class="s1">&#39;</span>
<span class="k">return</span> <span class="nb">abs</span><span class="p">(</span><span class="n">prevs_hat</span> <span class="o">-</span> <span class="n">prevs_true</span><span class="p">)</span><span class="o">.</span><span class="n">sum</span><span class="p">(</span><span class="n">axis</span><span class="o">=-</span><span class="mi">1</span><span class="p">)</span><span class="o">/</span><span class="p">(</span><span class="mi">2</span> <span class="o">*</span> <span class="p">(</span><span class="mi">1</span> <span class="o">-</span> <span class="n">prevs_true</span><span class="o">.</span><span class="n">min</span><span class="p">(</span><span class="n">axis</span><span class="o">=-</span><span class="mi">1</span><span class="p">)))</span></div>
<div class="viewcode-block" id="mnae">
<a class="viewcode-back" href="../../quapy.html#quapy.error.mnae">[docs]</a>
<span class="k">def</span><span class="w"> </span><span class="nf">mnae</span><span class="p">(</span><span class="n">prevs_true</span><span class="p">,</span> <span class="n">prevs_hat</span><span class="p">):</span>
<span class="w"> </span><span class="sd">&quot;&quot;&quot;Computes the mean normalized absolute error (see :meth:`quapy.error.nae`) across the sample pairs.</span>
<span class="sd"> :param prevs_true: array-like of shape `(n_samples, n_classes,)` with the true prevalence values</span>
<span class="sd"> :param prevs_hat: array-like of shape `(n_samples, n_classes,)` with the predicted</span>
<span class="sd"> prevalence values</span>
<span class="sd"> :return: mean normalized absolute error</span>
<span class="sd"> &quot;&quot;&quot;</span>
<span class="k">return</span> <span class="n">nae</span><span class="p">(</span><span class="n">prevs_true</span><span class="p">,</span> <span class="n">prevs_hat</span><span class="p">)</span><span class="o">.</span><span class="n">mean</span><span class="p">()</span></div>
<div class="viewcode-block" id="mse">
<a class="viewcode-back" href="../../quapy.html#quapy.error.mse">[docs]</a>
<span class="k">def</span><span class="w"> </span><span class="nf">mse</span><span class="p">(</span><span class="n">prevs_true</span><span class="p">,</span> <span class="n">prevs_hat</span><span class="p">):</span>
<span class="w"> </span><span class="sd">&quot;&quot;&quot;Computes the mean squared error (see :meth:`quapy.error.se`) across the sample pairs.</span>
<span class="sd"> :param prevs_true: array-like of shape `(n_samples, n_classes,)` with the</span>
<span class="sd"> true prevalence values</span>
<span class="sd"> :param prevs_hat: array-like of shape `(n_samples, n_classes,)` with the</span>
<span class="sd"> predicted prevalence values</span>
<span class="sd"> :return: mean squared error</span>
<span class="sd"> &quot;&quot;&quot;</span>
<span class="k">return</span> <span class="n">se</span><span class="p">(</span><span class="n">prevs_true</span><span class="p">,</span> <span class="n">prevs_hat</span><span class="p">)</span><span class="o">.</span><span class="n">mean</span><span class="p">()</span></div>
<div class="viewcode-block" id="se">
<a class="viewcode-back" href="../../quapy.html#quapy.error.se">[docs]</a>
<span class="k">def</span><span class="w"> </span><span class="nf">se</span><span class="p">(</span><span class="n">prevs_true</span><span class="p">,</span> <span class="n">prevs_hat</span><span class="p">):</span>
<span class="w"> </span><span class="sd">&quot;&quot;&quot;Computes the squared error between the two prevalence vectors.</span>
<span class="sd"> Squared error between two prevalence vectors :math:`p` and :math:`\\hat{p}` is computed as</span>
<span class="sd"> :math:`SE(p,\\hat{p})=\\frac{1}{|\\mathcal{Y}|}\\sum_{y\\in \\mathcal{Y}}(\\hat{p}(y)-p(y))^2`,</span>
<span class="sd"> where</span>
<span class="sd"> :math:`\\mathcal{Y}` are the classes of interest.</span>
<span class="sd"> :param prevs_true: array-like of shape `(n_classes,)` with the true prevalence values</span>
<span class="sd"> :param prevs_hat: array-like of shape `(n_classes,)` with the predicted prevalence values</span>
<span class="sd"> :return: absolute error</span>
<span class="sd"> &quot;&quot;&quot;</span>
<span class="n">prevs_true</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">asarray</span><span class="p">(</span><span class="n">prevs_true</span><span class="p">)</span>
<span class="n">prevs_hat</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">asarray</span><span class="p">(</span><span class="n">prevs_hat</span><span class="p">)</span>
<span class="k">return</span> <span class="p">((</span><span class="n">prevs_hat</span> <span class="o">-</span> <span class="n">prevs_true</span><span class="p">)</span> <span class="o">**</span> <span class="mi">2</span><span class="p">)</span><span class="o">.</span><span class="n">mean</span><span class="p">(</span><span class="n">axis</span><span class="o">=-</span><span class="mi">1</span><span class="p">)</span></div>
<div class="viewcode-block" id="sre">
<a class="viewcode-back" href="../../quapy.html#quapy.error.sre">[docs]</a>
<span class="k">def</span><span class="w"> </span><span class="nf">sre</span><span class="p">(</span><span class="n">prevs_true</span><span class="p">,</span> <span class="n">prevs_hat</span><span class="p">,</span> <span class="n">prevs_train</span><span class="p">,</span> <span class="n">eps</span><span class="o">=</span><span class="mf">0.</span><span class="p">):</span>
<span class="w"> </span><span class="sd">&quot;&quot;&quot;</span>
<span class="sd"> Computes the squared ratio error between two prevalence vectors.</span>
<span class="sd"> The squared ratio error between prevalence vectors :math:`p` and</span>
<span class="sd"> :math:`\\hat{p}` with training prevalence :math:`p^{tr}` is:</span>
<span class="sd"> :math:`SRE(p,\\hat{p},p^{tr})=\\frac{1}{|\\mathcal{Y}|}\\sum_{i \\in \\mathcal{Y}}(w_i-\\hat{w}_i)^2`,</span>
<span class="sd"> where :math:`w_i=\\frac{p_i}{p^{tr}_i}`.</span>
<span class="sd"> :param prevs_true: array-like with the true prevalence values</span>
<span class="sd"> :param prevs_hat: array-like with the predicted prevalence values</span>
<span class="sd"> :param prevs_train: array-like with the training prevalence values, or a single</span>
<span class="sd"> prevalence vector when all comparisons refer to the same training set</span>
<span class="sd"> :param eps: smoothing factor for the prevalence values (default 0, i.e., no smoothing)</span>
<span class="sd"> :return: squared ratio error</span>
<span class="sd"> &quot;&quot;&quot;</span>
<span class="n">prevs_true</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">asarray</span><span class="p">(</span><span class="n">prevs_true</span><span class="p">)</span>
<span class="n">prevs_hat</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">asarray</span><span class="p">(</span><span class="n">prevs_hat</span><span class="p">)</span>
<span class="n">prevs_train</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">asarray</span><span class="p">(</span><span class="n">prevs_train</span><span class="p">)</span>
<span class="k">assert</span> <span class="n">prevs_true</span><span class="o">.</span><span class="n">shape</span> <span class="o">==</span> <span class="n">prevs_hat</span><span class="o">.</span><span class="n">shape</span><span class="p">,</span> <span class="sa">f</span><span class="s1">&#39;wrong shape </span><span class="si">{</span><span class="n">prevs_true</span><span class="o">.</span><span class="n">shape</span><span class="si">=}</span><span class="s1"> vs </span><span class="si">{</span><span class="n">prevs_hat</span><span class="o">.</span><span class="n">shape</span><span class="si">=}</span><span class="s1">&#39;</span>
<span class="k">assert</span> <span class="n">prevs_true</span><span class="o">.</span><span class="n">shape</span><span class="p">[</span><span class="o">-</span><span class="mi">1</span><span class="p">]</span> <span class="o">==</span> <span class="n">prevs_train</span><span class="o">.</span><span class="n">shape</span><span class="p">[</span><span class="o">-</span><span class="mi">1</span><span class="p">],</span> <span class="s1">&#39;wrong shape for training prevalence&#39;</span>
<span class="k">if</span> <span class="n">prevs_true</span><span class="o">.</span><span class="n">ndim</span> <span class="o">==</span> <span class="mi">2</span> <span class="ow">and</span> <span class="n">prevs_train</span><span class="o">.</span><span class="n">ndim</span> <span class="o">==</span> <span class="mi">1</span><span class="p">:</span>
<span class="n">prevs_train</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">tile</span><span class="p">(</span><span class="n">prevs_train</span><span class="p">,</span> <span class="n">reps</span><span class="o">=</span><span class="p">(</span><span class="n">prevs_true</span><span class="o">.</span><span class="n">shape</span><span class="p">[</span><span class="mi">0</span><span class="p">],</span> <span class="mi">1</span><span class="p">))</span>
<span class="k">if</span> <span class="n">eps</span> <span class="o">&gt;</span> <span class="mi">0</span><span class="p">:</span>
<span class="n">prevs_true</span> <span class="o">=</span> <span class="n">smooth</span><span class="p">(</span><span class="n">prevs_true</span><span class="p">,</span> <span class="n">eps</span><span class="p">)</span>
<span class="n">prevs_hat</span> <span class="o">=</span> <span class="n">smooth</span><span class="p">(</span><span class="n">prevs_hat</span><span class="p">,</span> <span class="n">eps</span><span class="p">)</span>
<span class="n">prevs_train</span> <span class="o">=</span> <span class="n">smooth</span><span class="p">(</span><span class="n">prevs_train</span><span class="p">,</span> <span class="n">eps</span><span class="p">)</span>
<span class="n">n_classes</span> <span class="o">=</span> <span class="n">prevs_true</span><span class="o">.</span><span class="n">shape</span><span class="p">[</span><span class="o">-</span><span class="mi">1</span><span class="p">]</span>
<span class="n">w</span> <span class="o">=</span> <span class="n">prevs_true</span> <span class="o">/</span> <span class="n">prevs_train</span>
<span class="n">w_hat</span> <span class="o">=</span> <span class="n">prevs_hat</span> <span class="o">/</span> <span class="n">prevs_train</span>
<span class="k">return</span> <span class="p">(</span><span class="mf">1.</span> <span class="o">/</span> <span class="n">n_classes</span><span class="p">)</span> <span class="o">*</span> <span class="n">np</span><span class="o">.</span><span class="n">sum</span><span class="p">((</span><span class="n">w</span> <span class="o">-</span> <span class="n">w_hat</span><span class="p">)</span> <span class="o">**</span> <span class="mf">2.</span><span class="p">,</span> <span class="n">axis</span><span class="o">=-</span><span class="mi">1</span><span class="p">)</span></div>
<div class="viewcode-block" id="msre">
<a class="viewcode-back" href="../../quapy.html#quapy.error.msre">[docs]</a>
<span class="k">def</span><span class="w"> </span><span class="nf">msre</span><span class="p">(</span><span class="n">prevs_true</span><span class="p">,</span> <span class="n">prevs_hat</span><span class="p">,</span> <span class="n">prevs_train</span><span class="p">,</span> <span class="n">eps</span><span class="o">=</span><span class="mf">0.</span><span class="p">):</span>
<span class="w"> </span><span class="sd">&quot;&quot;&quot;</span>
<span class="sd"> Computes the mean squared ratio error (see :meth:`quapy.error.sre`) across the sample pairs.</span>
<span class="sd"> :param prevs_true: array-like of shape `(n_samples, n_classes,)` with the true prevalence values</span>
<span class="sd"> :param prevs_hat: array-like of shape equal to prevs_true with the predicted prevalence values</span>
<span class="sd"> :param prevs_train: array-like with the training prevalence values</span>
<span class="sd"> :param eps: smoothing factor (default 0, i.e., no smoothing)</span>
<span class="sd"> :return: mean squared ratio error</span>
<span class="sd"> &quot;&quot;&quot;</span>
<span class="k">return</span> <span class="n">np</span><span class="o">.</span><span class="n">mean</span><span class="p">(</span><span class="n">sre</span><span class="p">(</span><span class="n">prevs_true</span><span class="p">,</span> <span class="n">prevs_hat</span><span class="p">,</span> <span class="n">prevs_train</span><span class="p">,</span> <span class="n">eps</span><span class="p">))</span></div>
<div class="viewcode-block" id="aitchisondist">
<a class="viewcode-back" href="../../quapy.html#quapy.error.aitchisondist">[docs]</a>
<span class="k">def</span><span class="w"> </span><span class="nf">aitchisondist</span><span class="p">(</span><span class="n">prevs_true</span><span class="p">,</span> <span class="n">prevs_hat</span><span class="p">):</span>
<span class="w"> </span><span class="sd">&quot;&quot;&quot;</span>
<span class="sd"> Computes the Aitchison distance between two prevalence vectors.</span>
<span class="sd"> The Aitchison distance between prevalence vectors :math:`p` and</span>
<span class="sd"> :math:`\\hat{p}` is computed as</span>
<span class="sd"> :math:`d_A(p,\\hat{p})=\\|\\mathrm{clr}(p)-\\mathrm{clr}(\\hat{p})\\|_2`,</span>
<span class="sd"> where :math:`\\mathrm{clr}(p)_i=\\log p_i-\\frac{1}{|\\mathcal{Y}|}</span>
<span class="sd"> \\sum_{j \\in \\mathcal{Y}} \\log p_j`.</span>
<span class="sd"> :param prevs_true: array-like with the true prevalence values</span>
<span class="sd"> :param prevs_hat: array-like with the predicted prevalence values</span>
<span class="sd"> :return: Aitchison distance</span>
<span class="sd"> &quot;&quot;&quot;</span>
<span class="kn">from</span><span class="w"> </span><span class="nn">quapy.functional</span><span class="w"> </span><span class="kn">import</span> <span class="n">CLRtransformation</span>
<span class="n">clr</span> <span class="o">=</span> <span class="n">CLRtransformation</span><span class="p">()</span>
<span class="k">return</span> <span class="n">np</span><span class="o">.</span><span class="n">linalg</span><span class="o">.</span><span class="n">norm</span><span class="p">(</span><span class="n">clr</span><span class="p">(</span><span class="n">prevs_true</span><span class="p">)</span> <span class="o">-</span> <span class="n">clr</span><span class="p">(</span><span class="n">prevs_hat</span><span class="p">),</span> <span class="n">axis</span><span class="o">=-</span><span class="mi">1</span><span class="p">)</span></div>
<div class="viewcode-block" id="maitchisondist">
<a class="viewcode-back" href="../../quapy.html#quapy.error.maitchisondist">[docs]</a>
<span class="k">def</span><span class="w"> </span><span class="nf">maitchisondist</span><span class="p">(</span><span class="n">prevs_true</span><span class="p">,</span> <span class="n">prevs_hat</span><span class="p">):</span>
<span class="w"> </span><span class="sd">&quot;&quot;&quot;</span>
<span class="sd"> Computes the mean Aitchison distance (see :meth:`quapy.error.aitchisondist`)</span>
<span class="sd"> across the sample pairs, i.e.,</span>
<span class="sd"> :math:`\\mathrm{mAitchisonDist}=\\frac{1}{n}\\sum_{i=1}^n</span>
<span class="sd"> d_A(p_i,\\hat{p}_i)`.</span>
<span class="sd"> :param prevs_true: array-like with the true prevalence values</span>
<span class="sd"> :param prevs_hat: array-like with the predicted prevalence values</span>
<span class="sd"> :return: mean Aitchison distance</span>
<span class="sd"> &quot;&quot;&quot;</span>
<span class="k">return</span> <span class="n">np</span><span class="o">.</span><span class="n">mean</span><span class="p">(</span><span class="n">aitchisondist</span><span class="p">(</span><span class="n">prevs_true</span><span class="p">,</span> <span class="n">prevs_hat</span><span class="p">))</span></div>
<div class="viewcode-block" id="mkld">
<a class="viewcode-back" href="../../quapy.html#quapy.error.mkld">[docs]</a>
<span class="k">def</span><span class="w"> </span><span class="nf">mkld</span><span class="p">(</span><span class="n">prevs_true</span><span class="p">,</span> <span class="n">prevs_hat</span><span class="p">,</span> <span class="n">eps</span><span class="o">=</span><span class="kc">None</span><span class="p">):</span>
<span class="w"> </span><span class="sd">&quot;&quot;&quot;Computes the mean Kullback-Leibler divergence (see :meth:`quapy.error.kld`) across the</span>
<span class="sd"> sample pairs. The distributions are smoothed using the `eps` factor</span>
<span class="sd"> (see :meth:`quapy.error.smooth`).</span>
<span class="sd"> :param prevs_true: array-like of shape `(n_samples, n_classes,)` with the true</span>
<span class="sd"> prevalence values</span>
<span class="sd"> :param prevs_hat: array-like of shape `(n_samples, n_classes,)` with the predicted</span>
<span class="sd"> prevalence values</span>
<span class="sd"> :param eps: smoothing factor. KLD is not defined in cases in which the distributions contain</span>
<span class="sd"> zeros; `eps` is typically set to be :math:`\\frac{1}{2T}`, with :math:`T` the sample size.</span>
<span class="sd"> If `eps=None`, the sample size will be taken from the environment variable `SAMPLE_SIZE`</span>
<span class="sd"> (which has thus to be set beforehand).</span>
<span class="sd"> :return: mean Kullback-Leibler distribution</span>
<span class="sd"> &quot;&quot;&quot;</span>
<span class="k">return</span> <span class="n">kld</span><span class="p">(</span><span class="n">prevs_true</span><span class="p">,</span> <span class="n">prevs_hat</span><span class="p">,</span> <span class="n">eps</span><span class="p">)</span><span class="o">.</span><span class="n">mean</span><span class="p">()</span></div>
<div class="viewcode-block" id="kld">
<a class="viewcode-back" href="../../quapy.html#quapy.error.kld">[docs]</a>
<span class="k">def</span><span class="w"> </span><span class="nf">kld</span><span class="p">(</span><span class="n">prevs_true</span><span class="p">,</span> <span class="n">prevs_hat</span><span class="p">,</span> <span class="n">eps</span><span class="o">=</span><span class="kc">None</span><span class="p">):</span>
<span class="w"> </span><span class="sd">&quot;&quot;&quot;Computes the Kullback-Leibler divergence between the two prevalence distributions.</span>
<span class="sd"> Kullback-Leibler divergence between two prevalence distributions :math:`p` and :math:`\\hat{p}`</span>
<span class="sd"> is computed as</span>
<span class="sd"> :math:`KLD(p,\\hat{p})=D_{KL}(p||\\hat{p})=</span>
<span class="sd"> \\sum_{y\\in \\mathcal{Y}} p(y)\\log\\frac{p(y)}{\\hat{p}(y)}`,</span>
<span class="sd"> where :math:`\\mathcal{Y}` are the classes of interest.</span>
<span class="sd"> The distributions are smoothed using the `eps` factor (see :meth:`quapy.error.smooth`).</span>
<span class="sd"> :param prevs_true: array-like of shape `(n_classes,)` with the true prevalence values</span>
<span class="sd"> :param prevs_hat: array-like of shape `(n_classes,)` with the predicted prevalence values</span>
<span class="sd"> :param eps: smoothing factor. KLD is not defined in cases in which the distributions contain</span>
<span class="sd"> zeros; `eps` is typically set to be :math:`\\frac{1}{2T}`, with :math:`T` the sample size.</span>
<span class="sd"> If `eps=None`, the sample size will be taken from the environment variable `SAMPLE_SIZE`</span>
<span class="sd"> (which has thus to be set beforehand).</span>
<span class="sd"> :return: Kullback-Leibler divergence between the two distributions</span>
<span class="sd"> &quot;&quot;&quot;</span>
<span class="n">eps</span> <span class="o">=</span> <span class="n">__check_eps</span><span class="p">(</span><span class="n">eps</span><span class="p">)</span>
<span class="n">smooth_prevs</span> <span class="o">=</span> <span class="n">smooth</span><span class="p">(</span><span class="n">prevs_true</span><span class="p">,</span> <span class="n">eps</span><span class="p">)</span>
<span class="n">smooth_prevs_hat</span> <span class="o">=</span> <span class="n">smooth</span><span class="p">(</span><span class="n">prevs_hat</span><span class="p">,</span> <span class="n">eps</span><span class="p">)</span>
<span class="k">return</span> <span class="p">(</span><span class="n">smooth_prevs</span><span class="o">*</span><span class="n">np</span><span class="o">.</span><span class="n">log</span><span class="p">(</span><span class="n">smooth_prevs</span><span class="o">/</span><span class="n">smooth_prevs_hat</span><span class="p">))</span><span class="o">.</span><span class="n">sum</span><span class="p">(</span><span class="n">axis</span><span class="o">=-</span><span class="mi">1</span><span class="p">)</span></div>
<div class="viewcode-block" id="mnkld">
<a class="viewcode-back" href="../../quapy.html#quapy.error.mnkld">[docs]</a>
<span class="k">def</span><span class="w"> </span><span class="nf">mnkld</span><span class="p">(</span><span class="n">prevs_true</span><span class="p">,</span> <span class="n">prevs_hat</span><span class="p">,</span> <span class="n">eps</span><span class="o">=</span><span class="kc">None</span><span class="p">):</span>
<span class="w"> </span><span class="sd">&quot;&quot;&quot;Computes the mean Normalized Kullback-Leibler divergence (see :meth:`quapy.error.nkld`)</span>
<span class="sd"> across the sample pairs. The distributions are smoothed using the `eps` factor</span>
<span class="sd"> (see :meth:`quapy.error.smooth`).</span>
<span class="sd"> :param prevs_true: array-like of shape `(n_samples, n_classes,)` with the true prevalence values</span>
<span class="sd"> :param prevs_hat: array-like of shape `(n_samples, n_classes,)` with the predicted</span>
<span class="sd"> prevalence values</span>
<span class="sd"> :param eps: smoothing factor. NKLD is not defined in cases in which the distributions contain</span>
<span class="sd"> zeros; `eps` is typically set to be :math:`\\frac{1}{2T}`, with :math:`T` the sample size.</span>
<span class="sd"> If `eps=None`, the sample size will be taken from the environment variable `SAMPLE_SIZE`</span>
<span class="sd"> (which has thus to be set beforehand).</span>
<span class="sd"> :return: mean Normalized Kullback-Leibler distribution</span>
<span class="sd"> &quot;&quot;&quot;</span>
<span class="k">return</span> <span class="n">nkld</span><span class="p">(</span><span class="n">prevs_true</span><span class="p">,</span> <span class="n">prevs_hat</span><span class="p">,</span> <span class="n">eps</span><span class="p">)</span><span class="o">.</span><span class="n">mean</span><span class="p">()</span></div>
<div class="viewcode-block" id="nkld">
<a class="viewcode-back" href="../../quapy.html#quapy.error.nkld">[docs]</a>
<span class="k">def</span><span class="w"> </span><span class="nf">nkld</span><span class="p">(</span><span class="n">prevs_true</span><span class="p">,</span> <span class="n">prevs_hat</span><span class="p">,</span> <span class="n">eps</span><span class="o">=</span><span class="kc">None</span><span class="p">):</span>
<span class="w"> </span><span class="sd">&quot;&quot;&quot;Computes the Normalized Kullback-Leibler divergence between the two prevalence distributions.</span>
<span class="sd"> Normalized Kullback-Leibler divergence between two prevalence distributions :math:`p` and</span>
<span class="sd"> :math:`\\hat{p}` is computed as</span>
<span class="sd"> math:`NKLD(p,\\hat{p}) = 2\\frac{e^{KLD(p,\\hat{p})}}{e^{KLD(p,\\hat{p})}+1}-1`,</span>
<span class="sd"> where</span>
<span class="sd"> :math:`\\mathcal{Y}` are the classes of interest.</span>
<span class="sd"> The distributions are smoothed using the `eps` factor (see :meth:`quapy.error.smooth`).</span>
<span class="sd"> :param prevs_true: array-like of shape `(n_classes,)` with the true prevalence values</span>
<span class="sd"> :param prevs_hat: array-like of shape `(n_classes,)` with the predicted prevalence values</span>
<span class="sd"> :param eps: smoothing factor. NKLD is not defined in cases in which the distributions</span>
<span class="sd"> contain zeros; `eps` is typically set to be :math:`\\frac{1}{2T}`, with :math:`T` the sample</span>
<span class="sd"> size. If `eps=None`, the sample size will be taken from the environment variable</span>
<span class="sd"> `SAMPLE_SIZE` (which has thus to be set beforehand).</span>
<span class="sd"> :return: Normalized Kullback-Leibler divergence between the two distributions</span>
<span class="sd"> &quot;&quot;&quot;</span>
<span class="n">ekld</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">exp</span><span class="p">(</span><span class="n">kld</span><span class="p">(</span><span class="n">prevs_true</span><span class="p">,</span> <span class="n">prevs_hat</span><span class="p">,</span> <span class="n">eps</span><span class="p">))</span>
<span class="k">return</span> <span class="mf">2.</span> <span class="o">*</span> <span class="n">ekld</span> <span class="o">/</span> <span class="p">(</span><span class="mi">1</span> <span class="o">+</span> <span class="n">ekld</span><span class="p">)</span> <span class="o">-</span> <span class="mf">1.</span></div>
<div class="viewcode-block" id="mrae">
<a class="viewcode-back" href="../../quapy.html#quapy.error.mrae">[docs]</a>
<span class="k">def</span><span class="w"> </span><span class="nf">mrae</span><span class="p">(</span><span class="n">prevs_true</span><span class="p">,</span> <span class="n">prevs_hat</span><span class="p">,</span> <span class="n">eps</span><span class="o">=</span><span class="kc">None</span><span class="p">):</span>
<span class="w"> </span><span class="sd">&quot;&quot;&quot;Computes the mean relative absolute error (see :meth:`quapy.error.rae`) across</span>
<span class="sd"> the sample pairs. The distributions are smoothed using the `eps` factor (see</span>
<span class="sd"> :meth:`quapy.error.smooth`).</span>
<span class="sd"> :param prevs_true: array-like of shape `(n_samples, n_classes,)` with the true</span>
<span class="sd"> prevalence values</span>
<span class="sd"> :param prevs_hat: array-like of shape `(n_samples, n_classes,)` with the predicted</span>
<span class="sd"> prevalence values</span>
<span class="sd"> :param eps: smoothing factor. `mrae` is not defined in cases in which the true</span>
<span class="sd"> distribution contains zeros; `eps` is typically set to be :math:`\\frac{1}{2T}`,</span>
<span class="sd"> with :math:`T` the sample size. If `eps=None`, the sample size will be taken from</span>
<span class="sd"> the environment variable `SAMPLE_SIZE` (which has thus to be set beforehand).</span>
<span class="sd"> :return: mean relative absolute error</span>
<span class="sd"> &quot;&quot;&quot;</span>
<span class="k">return</span> <span class="n">rae</span><span class="p">(</span><span class="n">prevs_true</span><span class="p">,</span> <span class="n">prevs_hat</span><span class="p">,</span> <span class="n">eps</span><span class="p">)</span><span class="o">.</span><span class="n">mean</span><span class="p">()</span></div>
<div class="viewcode-block" id="rae">
<a class="viewcode-back" href="../../quapy.html#quapy.error.rae">[docs]</a>
<span class="k">def</span><span class="w"> </span><span class="nf">rae</span><span class="p">(</span><span class="n">prevs_true</span><span class="p">,</span> <span class="n">prevs_hat</span><span class="p">,</span> <span class="n">eps</span><span class="o">=</span><span class="kc">None</span><span class="p">):</span>
<span class="w"> </span><span class="sd">&quot;&quot;&quot;Computes the absolute relative error between the two prevalence vectors.</span>
<span class="sd"> Relative absolute error between two prevalence vectors :math:`p` and :math:`\\hat{p}`</span>
<span class="sd"> is computed as</span>
<span class="sd"> :math:`RAE(p,\\hat{p})=</span>
<span class="sd"> \\frac{1}{|\\mathcal{Y}|}\\sum_{y\\in \\mathcal{Y}}\\frac{|\\hat{p}(y)-p(y)|}{p(y)}`,</span>
<span class="sd"> where :math:`\\mathcal{Y}` are the classes of interest.</span>
<span class="sd"> The distributions are smoothed using the `eps` factor (see :meth:`quapy.error.smooth`).</span>
<span class="sd"> :param prevs_true: array-like of shape `(n_classes,)` with the true prevalence values</span>
<span class="sd"> :param prevs_hat: array-like of shape `(n_classes,)` with the predicted prevalence values</span>
<span class="sd"> :param eps: smoothing factor. `rae` is not defined in cases in which the true distribution</span>
<span class="sd"> contains zeros; `eps` is typically set to be :math:`\\frac{1}{2T}`, with :math:`T` the</span>
<span class="sd"> sample size. If `eps=None`, the sample size will be taken from the environment variable</span>
<span class="sd"> `SAMPLE_SIZE` (which has thus to be set beforehand).</span>
<span class="sd"> :return: relative absolute error</span>
<span class="sd"> &quot;&quot;&quot;</span>
<span class="n">eps</span> <span class="o">=</span> <span class="n">__check_eps</span><span class="p">(</span><span class="n">eps</span><span class="p">)</span>
<span class="n">prevs_true</span> <span class="o">=</span> <span class="n">smooth</span><span class="p">(</span><span class="n">prevs_true</span><span class="p">,</span> <span class="n">eps</span><span class="p">)</span>
<span class="n">prevs_hat</span> <span class="o">=</span> <span class="n">smooth</span><span class="p">(</span><span class="n">prevs_hat</span><span class="p">,</span> <span class="n">eps</span><span class="p">)</span>
<span class="k">return</span> <span class="p">(</span><span class="nb">abs</span><span class="p">(</span><span class="n">prevs_true</span> <span class="o">-</span> <span class="n">prevs_hat</span><span class="p">)</span> <span class="o">/</span> <span class="n">prevs_true</span><span class="p">)</span><span class="o">.</span><span class="n">mean</span><span class="p">(</span><span class="n">axis</span><span class="o">=-</span><span class="mi">1</span><span class="p">)</span></div>
<div class="viewcode-block" id="nrae">
<a class="viewcode-back" href="../../quapy.html#quapy.error.nrae">[docs]</a>
<span class="k">def</span><span class="w"> </span><span class="nf">nrae</span><span class="p">(</span><span class="n">prevs_true</span><span class="p">,</span> <span class="n">prevs_hat</span><span class="p">,</span> <span class="n">eps</span><span class="o">=</span><span class="kc">None</span><span class="p">):</span>
<span class="w"> </span><span class="sd">&quot;&quot;&quot;Computes the normalized absolute relative error between the two prevalence vectors.</span>
<span class="sd"> Relative absolute error between two prevalence vectors :math:`p` and :math:`\\hat{p}`</span>
<span class="sd"> is computed as</span>
<span class="sd"> :math:`NRAE(p,\\hat{p})= \\frac{RAE(p,\\hat{p})}{z_{RAE}}`,</span>
<span class="sd"> where</span>
<span class="sd"> :math:`z_{RAE} = \\frac{|\\mathcal{Y}|-1+\\frac{1-\\min_{y\\in \\mathcal{Y}} p(y)}{\\min_{y\\in \\mathcal{Y}} p(y)}}{|\\mathcal{Y}|}`</span>
<span class="sd"> and :math:`\\mathcal{Y}` are the classes of interest.</span>
<span class="sd"> The distributions are smoothed using the `eps` factor (see :meth:`quapy.error.smooth`).</span>
<span class="sd"> :param prevs_true: array-like of shape `(n_classes,)` with the true prevalence values</span>
<span class="sd"> :param prevs_hat: array-like of shape `(n_classes,)` with the predicted prevalence values</span>
<span class="sd"> :param eps: smoothing factor. `nrae` is not defined in cases in which the true distribution</span>
<span class="sd"> contains zeros; `eps` is typically set to be :math:`\\frac{1}{2T}`, with :math:`T` the</span>
<span class="sd"> sample size. If `eps=None`, the sample size will be taken from the environment variable</span>
<span class="sd"> `SAMPLE_SIZE` (which has thus to be set beforehand).</span>
<span class="sd"> :return: normalized relative absolute error</span>
<span class="sd"> &quot;&quot;&quot;</span>
<span class="n">eps</span> <span class="o">=</span> <span class="n">__check_eps</span><span class="p">(</span><span class="n">eps</span><span class="p">)</span>
<span class="n">prevs_true</span> <span class="o">=</span> <span class="n">smooth</span><span class="p">(</span><span class="n">prevs_true</span><span class="p">,</span> <span class="n">eps</span><span class="p">)</span>
<span class="n">prevs_hat</span> <span class="o">=</span> <span class="n">smooth</span><span class="p">(</span><span class="n">prevs_hat</span><span class="p">,</span> <span class="n">eps</span><span class="p">)</span>
<span class="n">min_p</span> <span class="o">=</span> <span class="n">prevs_true</span><span class="o">.</span><span class="n">min</span><span class="p">(</span><span class="n">axis</span><span class="o">=-</span><span class="mi">1</span><span class="p">)</span>
<span class="k">return</span> <span class="p">(</span><span class="nb">abs</span><span class="p">(</span><span class="n">prevs_true</span> <span class="o">-</span> <span class="n">prevs_hat</span><span class="p">)</span> <span class="o">/</span> <span class="n">prevs_true</span><span class="p">)</span><span class="o">.</span><span class="n">sum</span><span class="p">(</span><span class="n">axis</span><span class="o">=-</span><span class="mi">1</span><span class="p">)</span><span class="o">/</span><span class="p">(</span><span class="n">prevs_true</span><span class="o">.</span><span class="n">shape</span><span class="p">[</span><span class="o">-</span><span class="mi">1</span><span class="p">]</span> <span class="o">-</span> <span class="mi">1</span> <span class="o">+</span> <span class="p">(</span><span class="mi">1</span> <span class="o">-</span> <span class="n">min_p</span><span class="p">)</span> <span class="o">/</span> <span class="n">min_p</span><span class="p">)</span></div>
<div class="viewcode-block" id="mnrae">
<a class="viewcode-back" href="../../quapy.html#quapy.error.mnrae">[docs]</a>
<span class="k">def</span><span class="w"> </span><span class="nf">mnrae</span><span class="p">(</span><span class="n">prevs_true</span><span class="p">,</span> <span class="n">prevs_hat</span><span class="p">,</span> <span class="n">eps</span><span class="o">=</span><span class="kc">None</span><span class="p">):</span>
<span class="w"> </span><span class="sd">&quot;&quot;&quot;Computes the mean normalized relative absolute error (see :meth:`quapy.error.nrae`) across</span>
<span class="sd"> the sample pairs. The distributions are smoothed using the `eps` factor (see</span>
<span class="sd"> :meth:`quapy.error.smooth`).</span>
<span class="sd"> :param prevs_true: array-like of shape `(n_samples, n_classes,)` with the true</span>
<span class="sd"> prevalence values</span>
<span class="sd"> :param prevs_hat: array-like of shape `(n_samples, n_classes,)` with the predicted</span>
<span class="sd"> prevalence values</span>
<span class="sd"> :param eps: smoothing factor. `mnrae` is not defined in cases in which the true</span>
<span class="sd"> distribution contains zeros; `eps` is typically set to be :math:`\\frac{1}{2T}`,</span>
<span class="sd"> with :math:`T` the sample size. If `eps=None`, the sample size will be taken from</span>
<span class="sd"> the environment variable `SAMPLE_SIZE` (which has thus to be set beforehand).</span>
<span class="sd"> :return: mean normalized relative absolute error</span>
<span class="sd"> &quot;&quot;&quot;</span>
<span class="k">return</span> <span class="n">nrae</span><span class="p">(</span><span class="n">prevs_true</span><span class="p">,</span> <span class="n">prevs_hat</span><span class="p">,</span> <span class="n">eps</span><span class="p">)</span><span class="o">.</span><span class="n">mean</span><span class="p">()</span></div>
<div class="viewcode-block" id="nmd">
<a class="viewcode-back" href="../../quapy.html#quapy.error.nmd">[docs]</a>
<span class="k">def</span><span class="w"> </span><span class="nf">nmd</span><span class="p">(</span><span class="n">prevs_true</span><span class="p">,</span> <span class="n">prevs_hat</span><span class="p">):</span>
<span class="w"> </span><span class="sd">&quot;&quot;&quot;</span>
<span class="sd"> Computes the Normalized Match Distance; which is the Normalized Distance multiplied by the factor</span>
<span class="sd"> `1/(n-1)` to guarantee the measure ranges between 0 (best prediction) and 1 (worst prediction).</span>
<span class="sd"> :param prevs_true: array-like of shape `(n_classes,)` or `(n_instances, n_classes)` with the true prevalence values</span>
<span class="sd"> :param prevs_hat: array-like of shape `(n_classes,)` or `(n_instances, n_classes)` with the predicted prevalence values</span>
<span class="sd"> :return: float in [0,1]</span>
<span class="sd"> &quot;&quot;&quot;</span>
<span class="n">prevs_true</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">asarray</span><span class="p">(</span><span class="n">prevs_true</span><span class="p">)</span>
<span class="n">prevs_hat</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">asarray</span><span class="p">(</span><span class="n">prevs_hat</span><span class="p">)</span>
<span class="n">n</span> <span class="o">=</span> <span class="n">prevs_true</span><span class="o">.</span><span class="n">shape</span><span class="p">[</span><span class="o">-</span><span class="mi">1</span><span class="p">]</span>
<span class="k">return</span> <span class="p">(</span><span class="mf">1.</span><span class="o">/</span><span class="p">(</span><span class="n">n</span><span class="o">-</span><span class="mi">1</span><span class="p">))</span><span class="o">*</span><span class="n">np</span><span class="o">.</span><span class="n">mean</span><span class="p">(</span><span class="n">match_distance</span><span class="p">(</span><span class="n">prevs_true</span><span class="p">,</span> <span class="n">prevs_hat</span><span class="p">))</span></div>
<div class="viewcode-block" id="bias_binary">
<a class="viewcode-back" href="../../quapy.html#quapy.error.bias_binary">[docs]</a>
<span class="k">def</span><span class="w"> </span><span class="nf">bias_binary</span><span class="p">(</span><span class="n">prevs_true</span><span class="p">,</span> <span class="n">prevs_hat</span><span class="p">):</span>
<span class="w"> </span><span class="sd">&quot;&quot;&quot;</span>
<span class="sd"> Computes the (positive) bias in a binary problem. The bias is simply the difference between the</span>
<span class="sd"> predicted positive value and the true positive value, so that a positive such value indicates the</span>
<span class="sd"> prediction has positive bias (i.e., it tends to overestimate) the true value, and negative otherwise.</span>
<span class="sd"> :math:`bias(p,\\hat{p})=\\hat{p}_1-p_1`,</span>
<span class="sd"> :param prevs_true: array-like of shape `(n_samples, n_classes,)` with the true prevalence values</span>
<span class="sd"> :param prevs_hat: array-like of shape `(n_samples, n_classes,)` with the predicted</span>
<span class="sd"> prevalence values</span>
<span class="sd"> :return: binary bias</span>
<span class="sd"> &quot;&quot;&quot;</span>
<span class="n">prevs_true</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">asarray</span><span class="p">(</span><span class="n">prevs_true</span><span class="p">)</span>
<span class="n">prevs_hat</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">asarray</span><span class="p">(</span><span class="n">prevs_hat</span><span class="p">)</span>
<span class="k">assert</span> <span class="n">prevs_true</span><span class="o">.</span><span class="n">shape</span><span class="p">[</span><span class="o">-</span><span class="mi">1</span><span class="p">]</span> <span class="o">==</span> <span class="mi">2</span> <span class="ow">and</span> <span class="n">prevs_true</span><span class="o">.</span><span class="n">shape</span><span class="p">[</span><span class="o">-</span><span class="mi">1</span><span class="p">]</span> <span class="o">==</span> <span class="mi">2</span><span class="p">,</span> <span class="sa">f</span><span class="s1">&#39;bias_binary can only be applied to binary problems&#39;</span>
<span class="k">return</span> <span class="n">prevs_hat</span><span class="p">[</span><span class="o">...</span><span class="p">,</span><span class="mi">1</span><span class="p">]</span><span class="o">-</span><span class="n">prevs_true</span><span class="p">[</span><span class="o">...</span><span class="p">,</span><span class="mi">1</span><span class="p">]</span></div>
<div class="viewcode-block" id="mean_bias_binary">
<a class="viewcode-back" href="../../quapy.html#quapy.error.mean_bias_binary">[docs]</a>
<span class="k">def</span><span class="w"> </span><span class="nf">mean_bias_binary</span><span class="p">(</span><span class="n">prevs_true</span><span class="p">,</span> <span class="n">prevs_hat</span><span class="p">):</span>
<span class="w"> </span><span class="sd">&quot;&quot;&quot;</span>
<span class="sd"> Computes the mean of the (positive) bias in a binary problem.</span>
<span class="sd"> :param prevs_true: array-like of shape `(n_classes,)` with the true prevalence values</span>
<span class="sd"> :param prevs_hat: array-like of shape `(n_classes,)` with the predicted prevalence values</span>
<span class="sd"> :return: mean binary bias</span>
<span class="sd"> &quot;&quot;&quot;</span>
<span class="k">return</span> <span class="n">np</span><span class="o">.</span><span class="n">mean</span><span class="p">(</span><span class="n">bias_binary</span><span class="p">(</span><span class="n">prevs_true</span><span class="p">,</span> <span class="n">prevs_hat</span><span class="p">))</span></div>
<div class="viewcode-block" id="md">
<a class="viewcode-back" href="../../quapy.html#quapy.error.md">[docs]</a>
<span class="k">def</span><span class="w"> </span><span class="nf">md</span><span class="p">(</span><span class="n">prevs_true</span><span class="p">,</span> <span class="n">prevs_hat</span><span class="p">,</span> <span class="n">ERROR_TOL</span><span class="o">=</span><span class="mf">1E-3</span><span class="p">):</span>
<span class="w"> </span><span class="sd">&quot;&quot;&quot;</span>
<span class="sd"> Computes the Match Distance, under the assumption that the cost in mistaking class i with class i+1 is 1 in</span>
<span class="sd"> all cases.</span>
<span class="sd"> :param prevs_true: array-like of shape `(n_classes,)` or `(n_instances, n_classes)` with the true prevalence values</span>
<span class="sd"> :param prevs_hat: array-like of shape `(n_classes,)` or `(n_instances, n_classes)` with the predicted prevalence values</span>
<span class="sd"> :return: float</span>
<span class="sd"> &quot;&quot;&quot;</span>
<span class="n">P</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">cumsum</span><span class="p">(</span><span class="n">prevs_true</span><span class="p">,</span> <span class="n">axis</span><span class="o">=-</span><span class="mi">1</span><span class="p">)</span>
<span class="n">P_hat</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">cumsum</span><span class="p">(</span><span class="n">prevs_hat</span><span class="p">,</span> <span class="n">axis</span><span class="o">=-</span><span class="mi">1</span><span class="p">)</span>
<span class="k">assert</span> <span class="n">np</span><span class="o">.</span><span class="n">all</span><span class="p">(</span><span class="n">np</span><span class="o">.</span><span class="n">isclose</span><span class="p">(</span><span class="n">P_hat</span><span class="p">[</span><span class="o">...</span><span class="p">,</span> <span class="o">-</span><span class="mi">1</span><span class="p">],</span> <span class="mf">1.0</span><span class="p">,</span> <span class="n">rtol</span><span class="o">=</span><span class="n">ERROR_TOL</span><span class="p">)),</span> \
<span class="s1">&#39;arg error in match_distance: the array does not represent a valid distribution&#39;</span>
<span class="n">distances</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">abs</span><span class="p">(</span><span class="n">P</span><span class="o">-</span><span class="n">P_hat</span><span class="p">)</span>
<span class="k">return</span> <span class="n">distances</span><span class="p">[</span><span class="o">...</span><span class="p">,</span> <span class="p">:</span><span class="o">-</span><span class="mi">1</span><span class="p">]</span><span class="o">.</span><span class="n">sum</span><span class="p">(</span><span class="n">axis</span><span class="o">=-</span><span class="mi">1</span><span class="p">)</span></div>
<div class="viewcode-block" id="smooth">
<a class="viewcode-back" href="../../quapy.html#quapy.error.smooth">[docs]</a>
<span class="k">def</span><span class="w"> </span><span class="nf">smooth</span><span class="p">(</span><span class="n">prevs</span><span class="p">,</span> <span class="n">eps</span><span class="p">):</span>
<span class="w"> </span><span class="sd">&quot;&quot;&quot; Smooths a prevalence distribution with :math:`\\epsilon` (`eps`) as:</span>
<span class="sd"> :math:`\\underline{p}(y)=\\frac{\\epsilon+p(y)}{\\epsilon|\\mathcal{Y}|+</span>
<span class="sd"> \\displaystyle\\sum_{y\\in \\mathcal{Y}}p(y)}`</span>
<span class="sd"> :param prevs: array-like of shape `(n_classes,)` with the true prevalence values</span>
<span class="sd"> :param eps: smoothing factor</span>
<span class="sd"> :return: array-like of shape `(n_classes,)` with the smoothed distribution</span>
<span class="sd"> &quot;&quot;&quot;</span>
<span class="n">prevs</span> <span class="o">=</span> <span class="n">np</span><span class="o">.</span><span class="n">asarray</span><span class="p">(</span><span class="n">prevs</span><span class="p">)</span>
<span class="n">n_classes</span> <span class="o">=</span> <span class="n">prevs</span><span class="o">.</span><span class="n">shape</span><span class="p">[</span><span class="o">-</span><span class="mi">1</span><span class="p">]</span>
<span class="k">return</span> <span class="p">(</span><span class="n">prevs</span> <span class="o">+</span> <span class="n">eps</span><span class="p">)</span> <span class="o">/</span> <span class="p">(</span><span class="n">eps</span> <span class="o">*</span> <span class="n">n_classes</span> <span class="o">+</span> <span class="mi">1</span><span class="p">)</span></div>
<span class="k">def</span><span class="w"> </span><span class="nf">__check_eps</span><span class="p">(</span><span class="n">eps</span><span class="o">=</span><span class="kc">None</span><span class="p">):</span>
<span class="k">if</span> <span class="n">eps</span> <span class="ow">is</span> <span class="kc">None</span><span class="p">:</span>
<span class="n">sample_size</span> <span class="o">=</span> <span class="n">qp</span><span class="o">.</span><span class="n">environ</span><span class="p">[</span><span class="s1">&#39;SAMPLE_SIZE&#39;</span><span class="p">]</span>
<span class="k">if</span> <span class="n">sample_size</span> <span class="ow">is</span> <span class="kc">None</span><span class="p">:</span>
<span class="k">raise</span> <span class="ne">ValueError</span><span class="p">(</span><span class="s1">&#39;eps was not defined, and qp.environ[&quot;SAMPLE_SIZE&quot;] was not set&#39;</span><span class="p">)</span>
<span class="n">eps</span> <span class="o">=</span> <span class="mf">1.</span> <span class="o">/</span> <span class="p">(</span><span class="mf">2.</span> <span class="o">*</span> <span class="n">sample_size</span><span class="p">)</span>
<span class="k">return</span> <span class="n">eps</span>
<span class="n">CLASSIFICATION_ERROR</span> <span class="o">=</span> <span class="p">{</span><span class="n">f1e</span><span class="p">,</span> <span class="n">acce</span><span class="p">}</span>
<span class="n">QUANTIFICATION_ERROR</span> <span class="o">=</span> <span class="p">{</span><span class="n">mae</span><span class="p">,</span> <span class="n">mnae</span><span class="p">,</span> <span class="n">mrae</span><span class="p">,</span> <span class="n">mnrae</span><span class="p">,</span> <span class="n">mse</span><span class="p">,</span> <span class="n">mkld</span><span class="p">,</span> <span class="n">mnkld</span><span class="p">,</span> <span class="n">msre</span><span class="p">,</span> <span class="n">maitchisondist</span><span class="p">}</span>
<span class="n">QUANTIFICATION_ERROR_SINGLE</span> <span class="o">=</span> <span class="p">{</span><span class="n">ae</span><span class="p">,</span> <span class="n">nae</span><span class="p">,</span> <span class="n">rae</span><span class="p">,</span> <span class="n">nrae</span><span class="p">,</span> <span class="n">se</span><span class="p">,</span> <span class="n">kld</span><span class="p">,</span> <span class="n">nkld</span><span class="p">,</span> <span class="n">sre</span><span class="p">,</span> <span class="n">aitchisondist</span><span class="p">}</span>
<span class="n">QUANTIFICATION_ERROR_SMOOTH</span> <span class="o">=</span> <span class="p">{</span><span class="n">kld</span><span class="p">,</span> <span class="n">nkld</span><span class="p">,</span> <span class="n">rae</span><span class="p">,</span> <span class="n">nrae</span><span class="p">,</span> <span class="n">mkld</span><span class="p">,</span> <span class="n">mnkld</span><span class="p">,</span> <span class="n">mrae</span><span class="p">}</span>
<span class="n">CLASSIFICATION_ERROR_NAMES</span> <span class="o">=</span> <span class="p">{</span><span class="n">func</span><span class="o">.</span><span class="vm">__name__</span> <span class="k">for</span> <span class="n">func</span> <span class="ow">in</span> <span class="n">CLASSIFICATION_ERROR</span><span class="p">}</span>
<span class="n">QUANTIFICATION_ERROR_NAMES</span> <span class="o">=</span> <span class="p">{</span><span class="n">func</span><span class="o">.</span><span class="vm">__name__</span> <span class="k">for</span> <span class="n">func</span> <span class="ow">in</span> <span class="n">QUANTIFICATION_ERROR</span><span class="p">}</span>
<span class="n">QUANTIFICATION_ERROR_SINGLE_NAMES</span> <span class="o">=</span> <span class="p">{</span><span class="n">func</span><span class="o">.</span><span class="vm">__name__</span> <span class="k">for</span> <span class="n">func</span> <span class="ow">in</span> <span class="n">QUANTIFICATION_ERROR_SINGLE</span><span class="p">}</span>
<span class="n">QUANTIFICATION_ERROR_SMOOTH_NAMES</span> <span class="o">=</span> <span class="p">{</span><span class="n">func</span><span class="o">.</span><span class="vm">__name__</span> <span class="k">for</span> <span class="n">func</span> <span class="ow">in</span> <span class="n">QUANTIFICATION_ERROR_SMOOTH</span><span class="p">}</span>
<span class="n">ERROR_NAMES</span> <span class="o">=</span> \
<span class="n">CLASSIFICATION_ERROR_NAMES</span> <span class="o">|</span> <span class="n">QUANTIFICATION_ERROR_NAMES</span> <span class="o">|</span> <span class="n">QUANTIFICATION_ERROR_SINGLE_NAMES</span>
<span class="n">f1_error</span> <span class="o">=</span> <span class="n">f1e</span>
<span class="n">acc_error</span> <span class="o">=</span> <span class="n">acce</span>
<span class="n">mean_absolute_error</span> <span class="o">=</span> <span class="n">mae</span>
<span class="n">squared_ratio_error</span> <span class="o">=</span> <span class="n">sre</span>
<span class="n">dist_aitchison</span> <span class="o">=</span> <span class="n">aitchisondist</span>
<span class="n">mean_dist_aitchison</span> <span class="o">=</span> <span class="n">maitchisondist</span>
<span class="n">absolute_error</span> <span class="o">=</span> <span class="n">ae</span>
<span class="n">mean_relative_absolute_error</span> <span class="o">=</span> <span class="n">mrae</span>
<span class="n">relative_absolute_error</span> <span class="o">=</span> <span class="n">rae</span>
<span class="n">normalized_absolute_error</span> <span class="o">=</span> <span class="n">nae</span>
<span class="n">normalized_relative_absolute_error</span> <span class="o">=</span> <span class="n">nrae</span>
<span class="n">mean_normalized_absolute_error</span> <span class="o">=</span> <span class="n">mnae</span>
<span class="n">mean_normalized_relative_absolute_error</span> <span class="o">=</span> <span class="n">mnrae</span>
<span class="n">normalized_match_distance</span> <span class="o">=</span> <span class="n">nmd</span>
<span class="n">match_distance</span> <span class="o">=</span> <span class="n">md</span>
</pre></div>
</div>
</div>
<footer>
<hr/>
<div role="contentinfo">
<p>&#169; Copyright 2024, Alejandro Moreo.</p>
</div>
Built with <a href="https://www.sphinx-doc.org/">Sphinx</a> using a
<a href="https://github.com/readthedocs/sphinx_rtd_theme">theme</a>
provided by <a href="https://readthedocs.org">Read the Docs</a>.
</footer>
</div>
</div>
</section>
</div>
<script>
jQuery(function () {
SphinxRtdTheme.Navigation.enable(true);
});
</script>
</body>
</html>
Reference in New Issue View Git Blame Copy Permalink