{"id":6954,"date":"2025-09-16T11:14:11","date_gmt":"2025-09-16T09:14:11","guid":{"rendered":"https:\/\/francestat.com\/?page_id=6954"},"modified":"2026-03-30T11:37:30","modified_gmt":"2026-03-30T09:37:30","slug":"uniwin-svm","status":"publish","type":"page","link":"https:\/\/francestat.com\/index.php\/uniwin-svm\/","title":{"rendered":"UNIWIN – SVM"},"content":{"rendered":"

[vc_row][vc_column]

<\/div>

UNIWIN - S\u00e9parateurs \u00e0 vastes marges<\/h2><\/div><\/div>[\/vc_column][\/vc_row][vc_row][vc_column][vc_empty_space][\/vc_column][\/vc_row][vc_row][vc_column][vc_column_text]La proc\u00e9dure S\u00e9parateurs \u00e0 vastes marges<\/em> impl\u00e9mente une proc\u00e9dure d’apprentissage machine pour pr\u00e9voir des observations \u00e0 partir de donn\u00e9es. Elle cr\u00e9e des mod\u00e8les de deux formes : mod\u00e8les de classement<\/em> qui d\u00e9coupent des observations en groupes en se basant sur les caract\u00e9ristiques observ\u00e9es, mod\u00e8les de r\u00e9gression<\/em> qui pr\u00e9voient la valeur d’une variable cible.<\/p>\n

Dans le cas d’un mod\u00e8le de classement, l’algorithme d\u00e9coupent les observations en groupes en g\u00e9n\u00e9rant des marges autour des groupes aussi vastes que possible.<\/p>\n

Dans le cas d’un mod\u00e8le de r\u00e9gression, l’algorithme minimise les coefficients d’un mod\u00e8le dans lequel la distance des observations \u00e0 une r\u00e9gion autour du mod\u00e8le ajust\u00e9 d\u00e9finie par un montant d’erreur acceptable est aussi petite que possible.<\/p>\n

Les observations sont classiquement divis\u00e9es en trois jeux :<\/p>\n

    \n
  • Un jeu d’apprentissage<\/em> utilis\u00e9 pour construire le mod\u00e8le.<\/li>\n
  • Un jeu de validation<\/em>, pour lequel le groupe ou la valeur est connu, utilis\u00e9 pour valider le mod\u00e8le.<\/li>\n
  • Un jeu de pr\u00e9vision<\/em>, pour lequel le groupe ou la valeur n’est pas connu, utilis\u00e9 pour faire les pr\u00e9visions d\u00e9sir\u00e9es.<\/li>\n<\/ul>\n

    La variable \u00e0 expliquer est qualitative (classement) ou quantitative (r\u00e9gression).<\/p>\n

    Cette proc\u00e9dure est bas\u00e9e sur le package R ‘e1071’.[\/vc_column_text][\/vc_column][\/vc_row][vc_row][vc_column][vc_single_image image=\u00a0\u00bb6967″ img_size=\u00a0\u00bblarge\u00a0\u00bb alignment=\u00a0\u00bbcenter\u00a0\u00bb style=\u00a0\u00bbvc_box_border\u00a0\u00bb][\/vc_column][\/vc_row][vc_row][vc_column][vc_empty_space height=\u00a0\u00bb5px\u00a0\u00bb][vc_column_text]Tableaux<\/p>\n

    Param\u00e8tres optimaux des noyaux
    \nExactitude (classement, apprentissage)
    \nErreur quadratique (r\u00e9gression, apprentissage)
    \nImportance des variables
    \nVecteurs supports
    \nVecteurs supports par classe (classement)
    \nValeurs de d\u00e9cision (classement, apprentissage)
    \nPr\u00e9visions (classement, apprentissage)
    \nMatrice de confusion (classement, apprentissage)
    \nSensibilit\u00e9, sp\u00e9cificit\u00e9 (classement, apprentissage)
    \nValeurs de d\u00e9cision (classement, validation)
    \nPr\u00e9visions (classement, validation)
    \nMatrice de confusion (classement, validation)
    \nSensibilit\u00e9, sp\u00e9cificit\u00e9 (classement, validation)
    \nValeurs de d\u00e9cision (classement, pr\u00e9vision)
    \nPr\u00e9visions (classement, pr\u00e9vision)
    \nR\u00e9sultats apprentissage (r\u00e9gression)
    \nR\u00e9sultats valdiation (r\u00e9gression)
    \nR\u00e9sultats pr\u00e9vision (r\u00e9gression)<\/p>\n

    Graphiques<\/p>\n

    Importance des variables explicatives
    \nNuages de points (apprentissage, validation, pr\u00e9vision)
    \nGraphiques des fronti\u00e8res (apprentissage, validation)
    \nDaigrammes de la matrice de confusion (classement, apprentissage, validation)
    \nCourbes ROC (classement, apprentissage, validation)
    \nPr\u00e9vus vs observ\u00e9s (r\u00e9gression, apprentissage, validation)
    \nR\u00e9sidus vs observ\u00e9s (r\u00e9gression, apprentissage, validation)[\/vc_column_text][vc_empty_space height=\u00a0\u00bb5px\u00a0\u00bb][\/vc_column][\/vc_row][vc_row][vc_column][vc_btn title=\u00a0\u00bbConsulter la documentation compl\u00e8te\u00a0\u00bb align=\u00a0\u00bbcenter\u00a0\u00bb link=\u00a0\u00bburl:https%3A%2F%2Ffrancestat.com%2Ftelecharg%2FUniwin%2Fpdf%2FS%25e9parateurs%2520%25e0%2520vastes%2520marges.pdf|title:UNIWIN%20-%20SVM\u00a0\u00bb][\/vc_column][\/vc_row]<\/p>\n<\/div>","protected":false},"excerpt":{"rendered":"

    [vc_row][vc_column][\/vc_column][\/vc_row][vc_row][vc_column][vc_empty_space][\/vc_column][\/vc_row][vc_row][vc_column][vc_column_text]La proc\u00e9dure S\u00e9parateurs \u00e0 vastes marges impl\u00e9mente une proc\u00e9dure d’apprentissage machine pour pr\u00e9voir des observations \u00e0 partir de donn\u00e9es. Elle cr\u00e9e des mod\u00e8les de deux formes : mod\u00e8les de classement qui d\u00e9coupent des observations en groupes en se basant sur les caract\u00e9ristiques observ\u00e9es, mod\u00e8les de r\u00e9gression qui pr\u00e9voient la valeur d’une variable cible. Dans le…<\/p>\n","protected":false},"author":1,"featured_media":0,"parent":0,"menu_order":0,"comment_status":"closed","ping_status":"closed","template":"","meta":{"footnotes":""},"class_list":["post-6954","page","type-page","status-publish","hentry","description-off"],"yoast_head":"\nUNIWIN - SVM - FRANCESTAT<\/title>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/francestat.com\/index.php\/uniwin-svm\/\" \/>\n<meta property=\"og:locale\" content=\"fr_FR\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"UNIWIN - SVM - FRANCESTAT\" \/>\n<meta property=\"og:description\" content=\"[vc_row][vc_column][\/vc_column][\/vc_row][vc_row][vc_column][vc_empty_space][\/vc_column][\/vc_row][vc_row][vc_column][vc_column_text]La proc\u00e9dure S\u00e9parateurs \u00e0 vastes marges impl\u00e9mente une proc\u00e9dure d’apprentissage machine pour pr\u00e9voir des observations \u00e0 partir de donn\u00e9es. Elle cr\u00e9e des mod\u00e8les de deux formes : mod\u00e8les de classement qui d\u00e9coupent des observations en groupes en se basant sur les caract\u00e9ristiques observ\u00e9es, mod\u00e8les de r\u00e9gression qui pr\u00e9voient la valeur d’une variable cible. Dans le…\" \/>\n<meta property=\"og:url\" content=\"https:\/\/francestat.com\/index.php\/uniwin-svm\/\" \/>\n<meta property=\"og:site_name\" content=\"FRANCESTAT\" \/>\n<meta property=\"article:modified_time\" content=\"2026-03-30T09:37:30+00:00\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:label1\" content=\"Dur\u00e9e de lecture estim\u00e9e\" \/>\n\t<meta name=\"twitter:data1\" content=\"2 minutes\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\\\/\\\/schema.org\",\"@graph\":[{\"@type\":\"WebPage\",\"@id\":\"https:\\\/\\\/francestat.com\\\/index.php\\\/uniwin-svm\\\/\",\"url\":\"https:\\\/\\\/francestat.com\\\/index.php\\\/uniwin-svm\\\/\",\"name\":\"UNIWIN - SVM - FRANCESTAT\",\"isPartOf\":{\"@id\":\"https:\\\/\\\/francestat.com\\\/#website\"},\"datePublished\":\"2025-09-16T09:14:11+00:00\",\"dateModified\":\"2026-03-30T09:37:30+00:00\",\"breadcrumb\":{\"@id\":\"https:\\\/\\\/francestat.com\\\/index.php\\\/uniwin-svm\\\/#breadcrumb\"},\"inLanguage\":\"fr-FR\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\\\/\\\/francestat.com\\\/index.php\\\/uniwin-svm\\\/\"]}]},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\\\/\\\/francestat.com\\\/index.php\\\/uniwin-svm\\\/#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"Accueil\",\"item\":\"https:\\\/\\\/francestat.com\\\/\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"UNIWIN – SVM\"}]},{\"@type\":\"WebSite\",\"@id\":\"https:\\\/\\\/francestat.com\\\/#website\",\"url\":\"https:\\\/\\\/francestat.com\\\/\",\"name\":\"FRANCESTAT\",\"description\":\"Logiciels, formations et services statistiques\",\"publisher\":{\"@id\":\"https:\\\/\\\/francestat.com\\\/#organization\"},\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\\\/\\\/francestat.com\\\/?s={search_term_string}\"},\"query-input\":{\"@type\":\"PropertyValueSpecification\",\"valueRequired\":true,\"valueName\":\"search_term_string\"}}],\"inLanguage\":\"fr-FR\"},{\"@type\":\"Organization\",\"@id\":\"https:\\\/\\\/francestat.com\\\/#organization\",\"name\":\"FRANCESTAT\",\"url\":\"https:\\\/\\\/francestat.com\\\/\",\"logo\":{\"@type\":\"ImageObject\",\"inLanguage\":\"fr-FR\",\"@id\":\"https:\\\/\\\/francestat.com\\\/#\\\/schema\\\/logo\\\/image\\\/\",\"url\":\"https:\\\/\\\/francestat.com\\\/wp-content\\\/uploads\\\/2018\\\/05\\\/logo_horizontal_small_francestat.jpg\",\"contentUrl\":\"https:\\\/\\\/francestat.com\\\/wp-content\\\/uploads\\\/2018\\\/05\\\/logo_horizontal_small_francestat.jpg\",\"width\":155,\"height\":51,\"caption\":\"FRANCESTAT\"},\"image\":{\"@id\":\"https:\\\/\\\/francestat.com\\\/#\\\/schema\\\/logo\\\/image\\\/\"}}]}<\/script>\n<!-- \/ Yoast SEO plugin. -->","yoast_head_json":{"title":"UNIWIN - SVM - FRANCESTAT","robots":{"index":"index","follow":"follow","max-snippet":"max-snippet:-1","max-image-preview":"max-image-preview:large","max-video-preview":"max-video-preview:-1"},"canonical":"https:\/\/francestat.com\/index.php\/uniwin-svm\/","og_locale":"fr_FR","og_type":"article","og_title":"UNIWIN - SVM - FRANCESTAT","og_description":"[vc_row][vc_column][\/vc_column][\/vc_row][vc_row][vc_column][vc_empty_space][\/vc_column][\/vc_row][vc_row][vc_column][vc_column_text]La proc\u00e9dure S\u00e9parateurs \u00e0 vastes marges impl\u00e9mente une proc\u00e9dure d’apprentissage machine pour pr\u00e9voir des observations \u00e0 partir de donn\u00e9es. Elle cr\u00e9e des mod\u00e8les de deux formes : mod\u00e8les de classement qui d\u00e9coupent des observations en groupes en se basant sur les caract\u00e9ristiques observ\u00e9es, mod\u00e8les de r\u00e9gression qui pr\u00e9voient la valeur d’une variable cible. Dans le…","og_url":"https:\/\/francestat.com\/index.php\/uniwin-svm\/","og_site_name":"FRANCESTAT","article_modified_time":"2026-03-30T09:37:30+00:00","twitter_card":"summary_large_image","twitter_misc":{"Dur\u00e9e de lecture estim\u00e9e":"2 minutes"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"WebPage","@id":"https:\/\/francestat.com\/index.php\/uniwin-svm\/","url":"https:\/\/francestat.com\/index.php\/uniwin-svm\/","name":"UNIWIN - SVM - FRANCESTAT","isPartOf":{"@id":"https:\/\/francestat.com\/#website"},"datePublished":"2025-09-16T09:14:11+00:00","dateModified":"2026-03-30T09:37:30+00:00","breadcrumb":{"@id":"https:\/\/francestat.com\/index.php\/uniwin-svm\/#breadcrumb"},"inLanguage":"fr-FR","potentialAction":[{"@type":"ReadAction","target":["https:\/\/francestat.com\/index.php\/uniwin-svm\/"]}]},{"@type":"BreadcrumbList","@id":"https:\/\/francestat.com\/index.php\/uniwin-svm\/#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"Accueil","item":"https:\/\/francestat.com\/"},{"@type":"ListItem","position":2,"name":"UNIWIN – SVM"}]},{"@type":"WebSite","@id":"https:\/\/francestat.com\/#website","url":"https:\/\/francestat.com\/","name":"FRANCESTAT","description":"Logiciels, formations et services statistiques","publisher":{"@id":"https:\/\/francestat.com\/#organization"},"potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/francestat.com\/?s={search_term_string}"},"query-input":{"@type":"PropertyValueSpecification","valueRequired":true,"valueName":"search_term_string"}}],"inLanguage":"fr-FR"},{"@type":"Organization","@id":"https:\/\/francestat.com\/#organization","name":"FRANCESTAT","url":"https:\/\/francestat.com\/","logo":{"@type":"ImageObject","inLanguage":"fr-FR","@id":"https:\/\/francestat.com\/#\/schema\/logo\/image\/","url":"https:\/\/francestat.com\/wp-content\/uploads\/2018\/05\/logo_horizontal_small_francestat.jpg","contentUrl":"https:\/\/francestat.com\/wp-content\/uploads\/2018\/05\/logo_horizontal_small_francestat.jpg","width":155,"height":51,"caption":"FRANCESTAT"},"image":{"@id":"https:\/\/francestat.com\/#\/schema\/logo\/image\/"}}]}},"_links":{"self":[{"href":"https:\/\/francestat.com\/index.php\/wp-json\/wp\/v2\/pages\/6954","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/francestat.com\/index.php\/wp-json\/wp\/v2\/pages"}],"about":[{"href":"https:\/\/francestat.com\/index.php\/wp-json\/wp\/v2\/types\/page"}],"author":[{"embeddable":true,"href":"https:\/\/francestat.com\/index.php\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/francestat.com\/index.php\/wp-json\/wp\/v2\/comments?post=6954"}],"version-history":[{"count":12,"href":"https:\/\/francestat.com\/index.php\/wp-json\/wp\/v2\/pages\/6954\/revisions"}],"predecessor-version":[{"id":7172,"href":"https:\/\/francestat.com\/index.php\/wp-json\/wp\/v2\/pages\/6954\/revisions\/7172"}],"wp:attachment":[{"href":"https:\/\/francestat.com\/index.php\/wp-json\/wp\/v2\/media?parent=6954"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}