Support vector machines are a set of algorithms that learn from data by creating models that maximize their margin of error.<\/p>\n
Support vector machines (opens in new tab)<\/span><\/a>\u00a0(SVMs) are a family of algorithms for\u00a0classification (opens in new tab)<\/span><\/a>,\u00a0regression (opens in new tab)<\/span><\/a>,\u00a0transduction (opens in new tab)<\/span><\/a>, novelty detection (opens in new tab)<\/span><\/a>, and\u00a0semi-supervised learning (opens in new tab)<\/span><\/a>. They work by choosing a model that\u00a0maximizes the error margin (opens in new tab)<\/span><\/a> of a training set.<\/p>\n SVMs\u00a0were originally developed by\u00a0Vladimir Vapnik (opens in new tab)<\/span><\/a> in 1963. Since the mid-90s, a energetic research community has grown around them. If you want to learn more about SVMs, you can read Chris Burges’ tutorial<\/a>.\u00a0Nello Cristianini (opens in new tab)<\/span><\/a> and\u00a0John Shawe-Taylor (opens in new tab)<\/span><\/a> have written\u00a0a textbook (opens in new tab)<\/span><\/a> about them.\u00a0Bernhard Sch\u00f6lkopf (opens in new tab)<\/span><\/a> and\u00a0Alex Smola (opens in new tab)<\/span><\/a> wrote a textbook about kernel methods (opens in new tab)<\/span><\/a>, which are a closely-related set of methods.<\/p>\n Since 1998, we’ve done basic research into making SVMs be more user-friendly. Our research has resulted in:<\/p>\n See the list of publications, below, for complete citations.<\/p>\n\t The real-world data sets described in the technical report (below) are available in a compressed ASCII format (zip format). Both the\u00a0adult data and the\u00a0web data are available. There is a readme.txt file in each zip archive that explains the format of the file. The testing set for the adult data, the testing set for the web data set, and the\u00a0MNIST data set (opens in new tab)<\/span><\/a> is also available.<\/p>\n MSR currently does not have any software that implements SVMs.\u00a0LIBSVM (opens in new tab)<\/span><\/a> is a popular package that is based on a SMO-like algorithm.<\/p>\n Check here<\/a>\u00a0for errata on the SMO “Fast training”\u00a0physical\u00a0paper (already corrected in the on-line version).<\/p>\n\t\t\t<\/div>\n\t\t<\/div>\n\t<\/li>\n\t\t Sathiya Keerthi (opens in new tab)<\/span><\/a>\u00a0and colleagues have a\u00a0paper (opens in new tab)<\/span><\/a> that describes an improved SMO: instead of updating a single threshold, they update the bounds on permissible thresholds. They report substantial improvement in speed, especially for extreme C values.<\/p>\n Gary Flake<\/a>\u00a0and Steve Lawrence have an efficient SMO algorithm for Support Vector Regression<\/a>.<\/p>\n\t\t\t<\/div>\n\t\t<\/div>\n\t<\/li>\n\t\t There are some unclear ideas and errors in the \u201cFast Training\u201d paper that I (John Platt) would like to clarify on this web page:<\/p>\n Support vector machines are a set of algorithms that learn from data by creating models that maximize their margin of error. Support vector machines\u00a0(SVMs) are a family of algorithms for\u00a0classification,\u00a0regression,\u00a0transduction, novelty detection, and\u00a0semi-supervised learning. They work by choosing a model that\u00a0maximizes the error margin of a training set. SVMs\u00a0were originally developed by\u00a0Vladimir Vapnik in 1963. […]<\/p>\n","protected":false},"featured_media":0,"template":"","meta":{"msr-url-field":"","msr-podcast-episode":"","msrModifiedDate":"","msrModifiedDateEnabled":false,"ep_exclude_from_search":false,"_classifai_error":"","footnotes":""},"research-area":[13561,13556,13555],"msr-locale":[268875],"msr-impact-theme":[],"msr-pillar":[],"class_list":["post-169805","msr-project","type-msr-project","status-publish","hentry","msr-research-area-algorithms","msr-research-area-artificial-intelligence","msr-research-area-search-information-retrieval","msr-locale-en_us","msr-archive-status-active"],"msr_project_start":"2001-11-05","related-publications":[144958,151099,151241,152322,152409],"related-downloads":[],"related-videos":[],"related-groups":[],"related-events":[],"related-opportunities":[],"related-posts":[],"related-articles":[],"tab-content":[],"slides":[],"related-researchers":[],"msr_research_lab":[],"msr_impact_theme":[],"_links":{"self":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-project\/169805","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-project"}],"about":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/types\/msr-project"}],"version-history":[{"count":3,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-project\/169805\/revisions"}],"predecessor-version":[{"id":603543,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-project\/169805\/revisions\/603543"}],"wp:attachment":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/media?parent=169805"}],"wp:term":[{"taxonomy":"msr-research-area","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/research-area?post=169805"},{"taxonomy":"msr-locale","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-locale?post=169805"},{"taxonomy":"msr-impact-theme","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-impact-theme?post=169805"},{"taxonomy":"msr-pillar","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-pillar?post=169805"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}\n
\n\t\t\t\t\t\t\t\t
\n