GP/index.html

<!DOCTYPE html>
<html lang="en">

<head>
    <meta charset="UTF-8">
    <meta name="viewport" content="width=device-width, initial-scale=1.0">
    <title>Gaussian process visualization</title>
    
    <link rel="stylesheet" type="text/css" href="css/style.css">
    <link rel="stylesheet" href="./css/katex.min.css" integrity="sha384-zB1R0rpPzHqg7Kpt0Aljp8JPLqbXI3bhnPWROx27a9N0Ll6ZP/+DiW/UqRcLbRjq" crossorigin="anonymous">
    <script defer src="./js/katex.min.js" integrity="sha384-y23I5Q6l+B6vatafAwxRu/0oK/79VlbSz7Q9aiSZUvyWYIYsd+qj+o24G5ZU2zJz" crossorigin="anonymous"></script>
    <script src="./js/d3.v4.min.js"></script>    
    <script src="./js/numeric-1.2.6.js"></script>
</head>

<body>


    <header>
      <h1><a href="http://smlbook.org">The Supervised Machine Learning Book</a></h1>
    </header>

    <h2>Online material: Visualization of Gaussian Process regression</h2>
    <div style="font-size: 12px; margin-bottom: 20px">written by <a href="http://www.it.uu.se/katalog/johwa152/main">Johan W&aring;gberg</a> (at Uppsala University) 2019. The technical idea on how to smoothly loop over Gaussian process samples (as done in this animation) is described <a href="http://mlss.tuebingen.mpg.de/2013/Hennig_2013_Animating_Samples_from_Gaussian_Distributions.pdf">by Philipp Hennig (at University of T&uuml;bingen) in this document</a>.</div>
    
    <div style="margin-bottom: 50px">Try different kernels, change hyperparameters and add observations by clicking in the figure!</div>
    
     
    <div id="gp-viz"></div>
    <script type="module" src="index.js"></script>
</body>

</html>