
<!DOCTYPE html>

<html>

  <head>


    <meta http-equiv="content-type" content="text/html; charset=UTF-8">

  </head>

  <body>

    <div class="cooked">

      <p>Dear scikit-learn community</p>

      <p>I am very glad to announce the release of <a

          href="https://lorentzenchr.github.io/model-diagnostics/"

          rel="noopener nofollow ugc">model-diagnostics</a> 1.2 with its

        new <code>plot_marginal</code> and <code>compute_marginal</code>.</p>

      <p>Model-diagnostics helps you to assess calibration and

        performance of most supervised model - be it machine learning or

        statistical - for point predictions (like the mean, a quantile,

        or the probability for binary classification). The focus is on

        visualization and user-friendliness while well backed by

        statistical theory.</p>

      <p>The new <code>plot_marginal</code>, for instance, gives a

        great overview of the calibration as well as the model effect by

        a single feature, see <a

href="https://github.com/lorentzenchr/model-diagnostics/releases"

          rel="noopener nofollow ugc">1.2 release notes</a>:<br>

        <img

src="https://discuss.scientific-python.org/uploads/default/original/1X/c30da7cd7c1b3cebc6e8ae2781974a73651907e6.png"

          alt="image" data-base62-sha1="rPwmxajPZoqVWclVX6afSwOlv3U"

          style="aspect-ratio: 465 / 341;" width="465" height="341"></p>

      <p>Best,</p>

      <p>Christian</p>

    </div>

    <p></p>

  </body>

</html>