Auto-WEKA 2.0: Automatic model selection and hyperparameter optimization in WEKA

<html> <head>  <script async src="https://www.googletagmanager.com/gtag/js?id=UA-131826476-1"></script> <script> window.dataLayer = window.dataLayer || []; function gtag(){dataLayer.push(arguments);} gtag('js', new Date()); gtag('config', 'UA-131826476-1'); </script> <meta http-equiv="Content-type" content="text/html;charset=UTF-8">  <link rel="icon" href="/img/favicon.ico"> <link rel="icon" type="image/png" href="/img/favicon-16x16.png"> <link rel="icon" type="image/png" href="/img/favicon-32x32.png"> <title>Auto-WEKA 2.0: Automatic model selection and hyperparameter optimization in WEKA</title> <link rel="alternate" type="application/rss+xml" href="/jmlr.xml" title="JMLR RSS"> <link rel="stylesheet" type="text/css" href="/style.css"> <style type="text/css"> . {font-family:verdana,helvetica,sans-serif} a {text-decoration:none;color:#3030a0} #fixed { position: absolute; top: 0; left: 0; width: 8em; height: 100%; } body > #fixed { position: fixed; } #content { margin-top: 1em; margin-left: 10em; margin-right: 0.5em; } img.jmlr { width: 7em; } img.rss { width: 2em; } ul li { margin-bottom: 0.5em; } </style>  <script type="text/x-mathjax-config"> MathJax.Hub.Config({tex2jax: {inlineMath: [['$','$']]}}); </script> <script type="text/javascript" src="https://cdn.mathjax.org/mathjax/latest/MathJax.js?config=TeX-AMS-MML_HTMLorMML"> </script>  <meta name="citation_title" content="Auto-WEKA 2.0: Automatic model selection and hyperparameter optimization in WEKA"> <meta name="citation_author" content="Lars Kotthoff"> <meta name="citation_author" content="Chris Thornton"> <meta name="citation_author" content="Holger H. Hoos"> <meta name="citation_author" content="Frank Hutter"> <meta name="citation_author" content="Kevin Leyton-Brown"> <meta name="citation_journal_title" content="Journal of Machine Learning Research"> <meta name="citation_volume" content="18"> <meta name="citation_issue" content="25"> <meta name="citation_firstpage" content="1"> <meta name="citation_lastpage" content="5"> <meta name="citation_pdf_url" content="http://jmlr.org/papers/volume18/16-261/16-261.pdf"> <meta name="citation_publication_date" content="2017"> <meta name="citation_public_url" content="http://jmlr.org/papers/v18/16-261.html"> <meta name="citation_abstract_html_url" content="http://jmlr.org/papers/v18/16-261.html"> <meta name="citation_issn" content="1533-7928"> </head> <body> <div id="fixed"> <br> <a align="right" href="/" target=_top><img align="right" class="jmlr" src="/img/jmlr.jpg" border="0"></a> <p><br><br> <p align="right"> <A href="/"> Home Page </A> <p align="right"> <A href="/papers"> Papers </A> <p align="right"> <A href="/author-info.html"> Submissions </A> <p align="right"> <A href="/news.html"> News </A>  <p align="right" > <A href="/editorial-board.html"> Editorial Board </A> <p align="right" > <A href="/special_issues/"> Special Issues </A> <p align="right"> <A href="/mloss">Open Source Software</A> <p align="right"> <A href="https://proceedings.mlr.press/"> Proceedings (PMLR)</A> <p align="right"> <A href="https://data.mlr.press/"> Data (DMLR) </A> <p align="right"> <A href="/tmlr"> Transactions (TMLR) </A> <p align="right"> <A href="/search-jmlr.html"> Search </A> <p align="right"> <A href="/stats.html">Statistics</A> <p align="right"> <A href="/manudb"> Login </A></p> <p align="right"> <A href="/faq.html">Frequently Asked Questions </A></p> <p align="right"> <A href="/contact.html"> Contact Us </A></p> <br><br> <p align="right"> <A href="/jmlr.xml"> <img src="/img/RSS.gif" class="rss" alt="RSS Feed"> </A> </div> <div id="content"> <h2> Auto-WEKA 2.0: Automatic model selection and hyperparameter optimization in WEKA </h2> <p><b><i>Lars Kotthoff, Chris Thornton, Holger H. Hoos, Frank Hutter, Kevin Leyton-Brown</i></b>; 18(25):1−5, 2017.</p> <h3>Abstract</h3> <p class="abstract"> WEKA is a widely used, open-source machine learning platform. Due to its intuitive interface, it is particularly popular with novice users. However, such users often find it hard to identify the best approach for their particular dataset among the many available. We describe the new version of <em>Auto-WEKA</em>, a system designed to help such users by automatically searching through the joint space of WEKA's learning algorithms and their respective hyperparameter settings to maximize performance, using a state-of-the-art Bayesian optimization method. Our new package is tightly integrated with WEKA, making it just as accessible to end users as any other learning algorithm. </p> <font color="gray"><p>[abs]</font>[<a id="pdf" target="_blank" href="/papers/volume18/16-261/16-261.pdf">pdf</a>][<a id="bib" href="/papers/v18/16-261.bib">bib</a>]        [<a href="https://github.com/automl/autoweka">code</a>] [<a href="http://www.cs.ubc.ca/labs/beta/Projects/autoweka/">webpage</a>] <table width="100%"> <tr> <td align="right"><font size="-1">&copy <a target="_top" href="https://www.jmlr.org">JMLR</a> 2017. (<a href="https://github.com/JmlrOrg/v18/tree/main/16-261">edit</a>, <a href="http://jmlr.org/beta/papers/v18/16-261.html">beta</a>) </td> </tr> </table> </div>  <a style="font-size: 0" rel="me" href="https://sigmoid.social/@jmlr">Mastodon</a> </body> </html>

CINXE.COM

Auto-WEKA 2.0: Automatic model selection and hyperparameter optimization in WEKA