{"id":1203,"date":"2020-03-25T13:59:54","date_gmt":"2020-03-25T13:59:54","guid":{"rendered":"https:\/\/datascientest.com\/?p=1203"},"modified":"2024-04-23T09:56:50","modified_gmt":"2024-04-23T09:56:50","slug":"comment-gerer-les-problemes-de-classification-desequilibree-partie-i","status":"publish","type":"post","link":"https:\/\/datascientest.com\/comment-gerer-les-problemes-de-classification-desequilibree-partie-i","title":{"rendered":"Gestion des proble\u0300mes de Classification de\u0301se\u0301quilibre\u0301e &#8211; Partie I"},"content":{"rendered":"\t\t<div data-elementor-type=\"wp-post\" data-elementor-id=\"1203\" class=\"elementor elementor-1203\" data-elementor-post-type=\"post\">\n\t\t\t\t\t\t\t\t\t<section class=\"elementor-section elementor-top-section elementor-element elementor-element-9ad7276 elementor-section-boxed elementor-section-height-default elementor-section-height-default\" data-id=\"9ad7276\" data-element_type=\"section\">\n\t\t\t\t\t\t<div class=\"elementor-container elementor-column-gap-default\">\n\t\t\t\t\t<div class=\"elementor-column elementor-col-100 elementor-top-column elementor-element elementor-element-84041d0\" data-id=\"84041d0\" data-element_type=\"column\">\n\t\t\t<div class=\"elementor-widget-wrap elementor-element-populated\">\n\t\t\t\t\t\t\t\t<div class=\"elementor-element elementor-element-74c39dd elementor-widget elementor-widget-heading\" data-id=\"74c39dd\" data-element_type=\"widget\" data-widget_type=\"heading.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t<style>\/*! elementor - v3.15.0 - 02-08-2023 *\/\n.elementor-heading-title{padding:0;margin:0;line-height:1}.elementor-widget-heading .elementor-heading-title[class*=elementor-size-]>a{color:inherit;font-size:inherit;line-height:inherit}.elementor-widget-heading .elementor-heading-title.elementor-size-small{font-size:15px}.elementor-widget-heading .elementor-heading-title.elementor-size-medium{font-size:19px}.elementor-widget-heading .elementor-heading-title.elementor-size-large{font-size:29px}.elementor-widget-heading .elementor-heading-title.elementor-size-xl{font-size:39px}.elementor-widget-heading .elementor-heading-title.elementor-size-xxl{font-size:59px}<\/style><h2 class=\"elementor-heading-title elementor-size-default\">Comment g\u00e9rer les probl\u00e8mes de Classification d\u00e9s\u00e9quilibr\u00e9e ? <br>Partie I<\/h2>\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t<section class=\"elementor-section elementor-top-section elementor-element elementor-element-1526269 elementor-section-boxed elementor-section-height-default elementor-section-height-default\" data-id=\"1526269\" data-element_type=\"section\">\n\t\t\t\t\t\t<div class=\"elementor-container elementor-column-gap-default\">\n\t\t\t\t\t<div class=\"elementor-column elementor-col-100 elementor-top-column elementor-element elementor-element-7623a3f\" data-id=\"7623a3f\" data-element_type=\"column\">\n\t\t\t<div class=\"elementor-widget-wrap elementor-element-populated\">\n\t\t\t\t\t\t\t\t<div class=\"elementor-element elementor-element-2ed906c elementor-widget elementor-widget-text-editor\" data-id=\"2ed906c\" data-element_type=\"widget\" data-widget_type=\"text-editor.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t<style>\/*! elementor - v3.15.0 - 02-08-2023 *\/\n.elementor-widget-text-editor.elementor-drop-cap-view-stacked .elementor-drop-cap{background-color:#69727d;color:#fff}.elementor-widget-text-editor.elementor-drop-cap-view-framed .elementor-drop-cap{color:#69727d;border:3px solid;background-color:transparent}.elementor-widget-text-editor:not(.elementor-drop-cap-view-default) .elementor-drop-cap{margin-top:8px}.elementor-widget-text-editor:not(.elementor-drop-cap-view-default) .elementor-drop-cap-letter{width:1em;height:1em}.elementor-widget-text-editor .elementor-drop-cap{float:left;text-align:center;line-height:1;font-size:50px}.elementor-widget-text-editor .elementor-drop-cap-letter{display:inline-block}<\/style>\t\t\t\t<p>La classification sur donn\u00e9es d\u00e9s\u00e9quilibr\u00e9es est un probl\u00e8me de classification o\u00f9 l&rsquo;\u00e9chantillon d&rsquo;apprentissage contient une forte disparit\u00e9 entre les classes \u00e0 pr\u00e9dire. Ce probl\u00e8me revient fr\u00e9quemment dans les probl\u00e8mes de classification binaire, et notamment la d\u00e9tection d&rsquo;anomalies.<\/p>\n<p>\u00a0<span style=\"font-size: 14px;\">Cet article sera divis\u00e9 en deux parties : La premi\u00e8re s&rsquo;int\u00e9resse au choix des m\u00e9triques propres \u00e0 ce type de donn\u00e9es,\u00a0<\/span><a href=\"https:\/\/datascientest.com\/comment-gerer-les-problemes-de-classification-desequilibree-partie-ii\" target=\"_blank\" rel=\"noopener\"><u>la seconde<\/u><\/a><span style=\"font-size: 14px;\">\u00a0\u00a0d\u00e9taille l\u2019\u00e9ventail de m\u00e9thodes utiles pour obtenir un mod\u00e8le performant.\u00a0<\/span><\/p>\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t<section class=\"elementor-section elementor-top-section elementor-element elementor-element-3b65b47 elementor-section-boxed elementor-section-height-default elementor-section-height-default\" data-id=\"3b65b47\" data-element_type=\"section\">\n\t\t\t\t\t\t<div class=\"elementor-container elementor-column-gap-default\">\n\t\t\t\t\t<div class=\"elementor-column elementor-col-100 elementor-top-column elementor-element elementor-element-38e24fb\" data-id=\"38e24fb\" data-element_type=\"column\">\n\t\t\t<div class=\"elementor-widget-wrap elementor-element-populated\">\n\t\t\t\t\t\t\t\t<div class=\"elementor-element elementor-element-289bca8 elementor-widget elementor-widget-text-editor\" data-id=\"289bca8\" data-element_type=\"widget\" data-widget_type=\"text-editor.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t\t\t\t\t<h2>Partie I : Le choix des bonnes m\u00e9triques<\/h2>\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t<section class=\"elementor-section elementor-top-section elementor-element elementor-element-ea6a3e0 elementor-section-boxed elementor-section-height-default elementor-section-height-default\" data-id=\"ea6a3e0\" data-element_type=\"section\">\n\t\t\t\t\t\t<div class=\"elementor-container elementor-column-gap-default\">\n\t\t\t\t\t<div class=\"elementor-column elementor-col-100 elementor-top-column elementor-element elementor-element-e260851\" data-id=\"e260851\" data-element_type=\"column\">\n\t\t\t<div class=\"elementor-widget-wrap elementor-element-populated\">\n\t\t\t\t\t\t\t\t<div class=\"elementor-element elementor-element-4f17b7c elementor-widget elementor-widget-text-editor\" data-id=\"4f17b7c\" data-element_type=\"widget\" data-widget_type=\"text-editor.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t\t\t\t\t<h3><span style=\"font-size: 14px;\">Qu\u2019est ce qu\u2019une m\u00e9trique d\u2019\u00e9valuation ?\u00a0<\/span><\/h3>\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t<section class=\"elementor-section elementor-top-section elementor-element elementor-element-83a97e7 elementor-section-boxed elementor-section-height-default elementor-section-height-default\" data-id=\"83a97e7\" data-element_type=\"section\">\n\t\t\t\t\t\t<div class=\"elementor-container elementor-column-gap-default\">\n\t\t\t\t\t<div class=\"elementor-column elementor-col-100 elementor-top-column elementor-element elementor-element-8bd1e85\" data-id=\"8bd1e85\" data-element_type=\"column\">\n\t\t\t<div class=\"elementor-widget-wrap elementor-element-populated\">\n\t\t\t\t\t\t\t\t<div class=\"elementor-element elementor-element-21bf9aa elementor-widget elementor-widget-text-editor\" data-id=\"21bf9aa\" data-element_type=\"widget\" data-widget_type=\"text-editor.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t\t\t\t\t<p>Une m\u00e9trique d&rsquo;\u00e9valuation quantifie la performance d&rsquo;un mod\u00e8le pr\u00e9dictif.<\/p><p>Le choix de la bonne m\u00e9trique est donc crucial lors de l&rsquo;\u00e9valuation des mod\u00e8les de Machine Learning, et la qualit\u00e9 d&rsquo;un mod\u00e8le de classification d\u00e9pend directement de la m\u00e9trique utilis\u00e9e pour l&rsquo;\u00e9valuer.<\/p><p>Pour les probl\u00e8mes de classification, les m\u00e9triques consistent globalement \u00e0 comparer les classes r\u00e9elles aux classes pr\u00e9dites par le mod\u00e8le. Elles peuvent \u00e9galement permettre d\u2019interpr\u00e9ter les probabilit\u00e9s pr\u00e9dites pour ces classes.<\/p><p>L\u2019un des concepts cl\u00e9s de performance pour la classification est <a href=\"https:\/\/datascientest.com\/matrice-de-confusion\">la matrice de confusion<\/a> qui est une visualisation, sous forme de tableau, des pr\u00e9dictions du mod\u00e8le par rapport aux vrais labels. Chaque ligne de la matrice de confusion repr\u00e9sente les instances d&rsquo;une classe r\u00e9elles et chaque colonne repr\u00e9sente les instances d&rsquo;une classe pr\u00e9dite.<\/p><p>Prenons l\u2019exemple d\u2019une classification binaire, o\u00f9 l\u2019on dispose de 100 instances positives et 70 instances n\u00e9gatives.<\/p><p>La matrice de confusion ci-dessous correspond aux r\u00e9sultats obtenus par notre mod\u00e8le :<\/p><div>\u00a0<\/div>\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t<section class=\"elementor-section elementor-top-section elementor-element elementor-element-7d438dd elementor-section-boxed elementor-section-height-default elementor-section-height-default\" data-id=\"7d438dd\" data-element_type=\"section\">\n\t\t\t\t\t\t<div class=\"elementor-container elementor-column-gap-default\">\n\t\t\t\t\t<div class=\"elementor-column elementor-col-100 elementor-top-column elementor-element elementor-element-e115d14\" data-id=\"e115d14\" data-element_type=\"column\">\n\t\t\t<div class=\"elementor-widget-wrap elementor-element-populated\">\n\t\t\t\t\t\t\t\t<div class=\"elementor-element elementor-element-5c00f3e elementor-widget elementor-widget-image\" data-id=\"5c00f3e\" data-element_type=\"widget\" data-widget_type=\"image.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t<style>\/*! elementor - v3.15.0 - 02-08-2023 *\/\n.elementor-widget-image{text-align:center}.elementor-widget-image a{display:inline-block}.elementor-widget-image a img[src$=\".svg\"]{width:48px}.elementor-widget-image img{vertical-align:middle;display:inline-block}<\/style>\t\t\t\t\t\t\t\t\t\t\t\t<img decoding=\"async\" width=\"512\" height=\"288\" src=\"https:\/\/datascientest.com\/wp-content\/uploads\/2020\/03\/Image1-.png\" class=\"attachment-large size-large wp-image-194158\" alt=\"Matrice de confusion\" srcset=\"https:\/\/datascientest.com\/wp-content\/uploads\/2020\/03\/Image1-.png 512w, https:\/\/datascientest.com\/wp-content\/uploads\/2020\/03\/Image1--300x169.png 300w\" sizes=\"(max-width: 512px) 100vw, 512px\" \/>\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t<section class=\"elementor-section elementor-top-section elementor-element elementor-element-2813141 elementor-section-boxed elementor-section-height-default elementor-section-height-default\" data-id=\"2813141\" data-element_type=\"section\">\n\t\t\t\t\t\t<div class=\"elementor-container elementor-column-gap-default\">\n\t\t\t\t\t<div class=\"elementor-column elementor-col-100 elementor-top-column elementor-element elementor-element-f14621f\" data-id=\"f14621f\" data-element_type=\"column\">\n\t\t\t<div class=\"elementor-widget-wrap elementor-element-populated\">\n\t\t\t\t\t\t\t\t<div class=\"elementor-element elementor-element-7e1ea21 elementor-widget elementor-widget-text-editor\" data-id=\"7e1ea21\" data-element_type=\"widget\" data-widget_type=\"text-editor.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t\t\t\t\t<p>Elle permet d\u2019obtenir une vue d\u2019ensemble des pr\u00e9dictions justes et des pr\u00e9dictions fausse.<\/p><p>Pour r\u00e9sumer cette matrice en une m\u00e9trique, il est possible d\u2019utiliser le taux de bonnes pr\u00e9dictions ou accuracy . Ici il est \u00e9gal \u00e0 (90+57)\/170 = 0.86.<\/p><p>Le choix d&rsquo;une m\u00e9trique appropri\u00e9e n\u2019est pas \u00e9vident pour tout mod\u00e8le de Machine Learning, mais il est particuli\u00e8rement difficile pour les probl\u00e8mes de classification d\u00e9s\u00e9quilibr\u00e9e.<\/p><p>Dans le cas de donn\u00e9es comportant une classe fortement majoritaire, les algorithmes classiques sont souvent biais\u00e9s car leurs fonctions de perte tentent d&rsquo;optimiser des m\u00e9triques telles que le taux de bonnes pr\u00e9dictions, en ne tenant pas compte de la distribution des donn\u00e9es.<\/p><p>Dans le pire des cas, les classes minoritaires sont trait\u00e9es comme des valeurs aberrantes de la classe majoritaire et l&rsquo;algorithme d&rsquo;apprentissage g\u00e9n\u00e8re simplement un classifieur trivial qui classe chaque exemple dans la classe majoritaire. Le mod\u00e8le semblera performant mais cela ne sera que le reflet de la surrepr\u00e9sentation de la classe majoritaire. On parle d&rsquo;accuracy paradoxe.<\/p><p>Dans la majorit\u00e9 des cas, c&rsquo;est justement la classe minoritaire qui pr\u00e9sente le plus fort int\u00e9r\u00eat et que l&rsquo;on aimerait pouvoir identifier, comme dans l&rsquo;exemple de la d\u00e9tection de fraudes.<\/p><p>Le degr\u00e9 de d\u00e9s\u00e9quilibre varie mais les cas d&rsquo;usages sont fr\u00e9quents: d\u00e9pistage de maladies, d\u00e9tection de panne, moteur de recherche, filtrage de spams, ciblage marketing etc\u2026<\/p><div>\u00a0<\/div>\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t<section class=\"elementor-section elementor-top-section elementor-element elementor-element-a7e75fb elementor-section-boxed elementor-section-height-default elementor-section-height-default\" data-id=\"a7e75fb\" data-element_type=\"section\">\n\t\t\t\t\t\t<div class=\"elementor-container elementor-column-gap-default\">\n\t\t\t\t\t<div class=\"elementor-column elementor-col-100 elementor-top-column elementor-element elementor-element-1f5d40f\" data-id=\"1f5d40f\" data-element_type=\"column\">\n\t\t\t<div class=\"elementor-widget-wrap elementor-element-populated\">\n\t\t\t\t\t\t\t\t<div class=\"elementor-element elementor-element-f6b7053 elementor-widget elementor-widget-text-editor\" data-id=\"f6b7053\" data-element_type=\"widget\" data-widget_type=\"text-editor.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t\t\t\t\t<h3><span style=\"font-size: 14px;\">Mise en pratique : Churn Rate\u00a0<\/span><\/h3>\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t<section class=\"elementor-section elementor-top-section elementor-element elementor-element-976d063 elementor-section-boxed elementor-section-height-default elementor-section-height-default\" data-id=\"976d063\" data-element_type=\"section\">\n\t\t\t\t\t\t<div class=\"elementor-container elementor-column-gap-default\">\n\t\t\t\t\t<div class=\"elementor-column elementor-col-100 elementor-top-column elementor-element elementor-element-b6dc42b\" data-id=\"b6dc42b\" data-element_type=\"column\">\n\t\t\t<div class=\"elementor-widget-wrap elementor-element-populated\">\n\t\t\t\t\t\t\t\t<div class=\"elementor-element elementor-element-aaec347 elementor-widget elementor-widget-text-editor\" data-id=\"aaec347\" data-element_type=\"widget\" data-widget_type=\"text-editor.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t\t\t\t\t<p>Supposons qu\u2019une entreprise de services souhaite pr\u00e9dire son taux d\u2019attrition : \u2018churn rate\u2019 en anglais.\u00a0<\/p><p>Pour rappel le churn rate correspond au ratio : nombre de clients perdus \/ nombre de clients total , mesur\u00e9 sur une p\u00e9riode donn\u00e9e qui est le plus souvent l&rsquo;ann\u00e9e.<\/p><p>Pour cela, elle souhaite pr\u00e9dire pour chaque client s\u2019il r\u00e9siliera son contrat \u00e0 la fin de l\u2019ann\u00e9e.<\/p><p>On dispose d\u2019un jeu de donn\u00e9es contenant des informations personnelles et des caract\u00e9ristiques relatives aux contrats de chaque client de l\u2019entreprise pour l\u2019ann\u00e9e X, ainsi qu\u2019une variable permettant de savoir s\u2019il a renouvel\u00e9 son contrat \u00e0 la fin de l\u2019ann\u00e9e.\u00a0<\/p><p>Dans nos donn\u00e9es, le nombre de \u2018churners\u2019 correspond \u00e0\u00a0 environ 11% du nombre total de clients.<\/p><p>On d\u00e9cide d\u2019entra\u00eener un premier mod\u00e8le de R\u00e9gression logistique sur nos donn\u00e9es pr\u00e9par\u00e9es et normalis\u00e9es.<\/p>\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-61de7a2 elementor-widget elementor-widget-shortcode\" data-id=\"61de7a2\" data-element_type=\"widget\" data-widget_type=\"shortcode.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t\t\t<div class=\"elementor-shortcode\"><script src=\"https:\/\/gist.github.com\/Raphkas15\/473a8d6ef5535da5e1c02cdfd1977030.js\"><\/script><\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t<section class=\"elementor-section elementor-top-section elementor-element elementor-element-9936546 elementor-section-boxed elementor-section-height-default elementor-section-height-default\" data-id=\"9936546\" data-element_type=\"section\">\n\t\t\t\t\t\t<div class=\"elementor-container elementor-column-gap-default\">\n\t\t\t\t\t<div class=\"elementor-column elementor-col-100 elementor-top-column elementor-element elementor-element-b3cba1c\" data-id=\"b3cba1c\" data-element_type=\"column\">\n\t\t\t<div class=\"elementor-widget-wrap elementor-element-populated\">\n\t\t\t\t\t\t\t\t<div class=\"elementor-element elementor-element-d849354 elementor-widget elementor-widget-text-editor\" data-id=\"d849354\" data-element_type=\"widget\" data-widget_type=\"text-editor.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t\t\t\t\t<p>Surprise ! Notre code affiche un taux de bonnes pr\u00e9dictions de 0.90 !<\/p>\n<p>&nbsp;<span style=\"font-size: 14px;\">C&rsquo;est un score tr\u00e8s correct, mais rappelons-nous notre objectif : R\u00e9ussir \u00e0 pr\u00e9dire les d\u00e9parts \u00e9ventuels des clients.&nbsp;<\/span><span style=\"font-size: 14px;\">Est-ce que ce r\u00e9sultat signifie que sur 10 clients \u00ab\u00a0churners\u00a0\u00bb, 9 seront identifi\u00e9s comme tels par le mod\u00e8le? Absolument pas !<\/span><\/p><p>La seule interpr\u00e9tation que l\u2019on puisse faire est que 9 clients sur 10 ont \u00e9t\u00e9 bien class\u00e9s pas le mod\u00e8le.<br><\/p>\n<p><span style=\"font-size: 14px;\">Pour r\u00e9ussir \u00e0 d\u00e9tecter le comportement na\u00eff d&rsquo;un mod\u00e8le, l&rsquo;outil le plus efficace est toujours la matrice de confusion.<\/span><\/p>\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t<section class=\"elementor-section elementor-top-section elementor-element elementor-element-5854752 elementor-section-boxed elementor-section-height-default elementor-section-height-default\" data-id=\"5854752\" data-element_type=\"section\">\n\t\t\t\t\t\t<div class=\"elementor-container elementor-column-gap-default\">\n\t\t\t\t\t<div class=\"elementor-column elementor-col-100 elementor-top-column elementor-element elementor-element-5a0b0d2\" data-id=\"5a0b0d2\" data-element_type=\"column\">\n\t\t\t<div class=\"elementor-widget-wrap elementor-element-populated\">\n\t\t\t\t\t\t\t\t<div class=\"elementor-element elementor-element-d35d367 elementor-widget elementor-widget-shortcode\" data-id=\"d35d367\" data-element_type=\"widget\" data-widget_type=\"shortcode.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t\t\t<div class=\"elementor-shortcode\"><script src=\"https:\/\/gist.github.com\/Raphkas15\/48118451ce0bfb203caf95c55dd6d8d8.js\"><\/script><\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t<section class=\"elementor-section elementor-top-section elementor-element elementor-element-707cb91 elementor-section-boxed elementor-section-height-default elementor-section-height-default\" data-id=\"707cb91\" data-element_type=\"section\">\n\t\t\t\t\t\t<div class=\"elementor-container elementor-column-gap-default\">\n\t\t\t\t\t<div class=\"elementor-column elementor-col-100 elementor-top-column elementor-element elementor-element-68b8e09\" data-id=\"68b8e09\" data-element_type=\"column\">\n\t\t\t<div class=\"elementor-widget-wrap elementor-element-populated\">\n\t\t\t\t\t\t\t\t<div class=\"elementor-element elementor-element-fbb5db9 elementor-widget elementor-widget-text-editor\" data-id=\"fbb5db9\" data-element_type=\"widget\" data-widget_type=\"text-editor.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t\t\t\t\t<p><span style=\"font-size: 14px;\">Un premier coup d&rsquo;oeil a la matrice de confusion nous indique que le bon taux de bonnes pr\u00e9dictions obtenu est largement influenc\u00e9 par le bon comportement du mod\u00e8le sur la classe dominante (0).<\/span><\/p><p>Afin d&rsquo;\u00e9valuer le mod\u00e8le par rapport au comportement souhait\u00e9 sur une classe, il est possible d&rsquo;utiliser une s\u00e9rie de m\u00e9triques issues de la matrice de confusion, comme notamment la <b>pr\u00e9cision<\/b>, le <b>rappel <\/b>(recall) et le <b>f1-score<\/b>, d\u00e9finis ci-dessous.<\/p>\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t<section class=\"elementor-section elementor-top-section elementor-element elementor-element-137d81b elementor-section-boxed elementor-section-height-default elementor-section-height-default\" data-id=\"137d81b\" data-element_type=\"section\">\n\t\t\t\t\t\t<div class=\"elementor-container elementor-column-gap-default\">\n\t\t\t\t\t<div class=\"elementor-column elementor-col-100 elementor-top-column elementor-element elementor-element-16dcebd\" data-id=\"16dcebd\" data-element_type=\"column\">\n\t\t\t<div class=\"elementor-widget-wrap elementor-element-populated\">\n\t\t\t\t\t\t\t\t<div class=\"elementor-element elementor-element-7cfdded elementor-widget elementor-widget-image\" data-id=\"7cfdded\" data-element_type=\"widget\" data-widget_type=\"image.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<img decoding=\"async\" width=\"512\" height=\"89\" src=\"https:\/\/datascientest.com\/wp-content\/uploads\/2020\/03\/Image3.png\" class=\"attachment-large size-large wp-image-194160\" alt=\"S\u00e9rie de donn\u00e9e\" srcset=\"https:\/\/datascientest.com\/wp-content\/uploads\/2020\/03\/Image3.png 512w, https:\/\/datascientest.com\/wp-content\/uploads\/2020\/03\/Image3-300x52.png 300w\" sizes=\"(max-width: 512px) 100vw, 512px\" \/>\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t<section class=\"elementor-section elementor-top-section elementor-element elementor-element-f04c82f elementor-section-boxed elementor-section-height-default elementor-section-height-default\" data-id=\"f04c82f\" data-element_type=\"section\">\n\t\t\t\t\t\t<div class=\"elementor-container elementor-column-gap-default\">\n\t\t\t\t\t<div class=\"elementor-column elementor-col-100 elementor-top-column elementor-element elementor-element-bc33339\" data-id=\"bc33339\" data-element_type=\"column\">\n\t\t\t<div class=\"elementor-widget-wrap elementor-element-populated\">\n\t\t\t\t\t\t\t\t<div class=\"elementor-element elementor-element-064ca3c elementor-widget elementor-widget-text-editor\" data-id=\"064ca3c\" data-element_type=\"widget\" data-widget_type=\"text-editor.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t\t\t\t\t<p>Ainsi pour une classe donn\u00e9e:<\/p><p>Une pr\u00e9cision et un rappel \u00e9lev\u00e9 -&gt; La classe a \u00e9t\u00e9 bien g\u00e9r\u00e9e par le mod\u00e8le<\/p><p>Une pr\u00e9cision \u00e9lev\u00e9e et un rappel faible -&gt; La classe n&rsquo;est pas bien d\u00e9tect\u00e9e mais lorsqu&rsquo;elle l&rsquo;est, le mod\u00e8le est tr\u00e8s fiable.\u00a0<\/p><p>Une pr\u00e9cision faible et un rappel \u00e9lev\u00e9 -&gt; La classe est bien d\u00e9tect\u00e9e, mais inclut \u00e9galement des observations d&rsquo;autres classes.<\/p><p>\u00a0Une pr\u00e9cision et un rappel faible -&gt; la classe n&rsquo;a pas du tout \u00e9t\u00e9 bien g\u00e9r\u00e9e<\/p><p>Le F1-score permet de mesurer la pr\u00e9cision et le rappel \u00e0 la fois.<\/p><p>Dans le cas d&rsquo;une classification binaire, la sensibilit\u00e9 et la sp\u00e9cificit\u00e9 correspondent respectivement aux rappels de la classe positive et de la classe n\u00e9gative.<\/p><p>Une autre m\u00e9trique , la <b>moyenne g\u00e9om\u00e9trique\u00a0(G-mean)<\/b><span style=\"font-size: 14px;\">\u00a0, s&rsquo;av\u00e8re utile pour les probl\u00e8mes de classification d\u00e9s\u00e9quilibr\u00e9e: il s&rsquo;agit de la racine du produit de la sensibilit\u00e9 et de la sp\u00e9cificit\u00e9.<\/span><\/p><p>Ces diff\u00e9rentes m\u00e9triques sont facilement accessibles gr\u00e2ce au package\u00a0<span style=\"text-decoration: underline;\">imblearn<\/span><span style=\"font-size: 14px;\">.\u00a0<\/span><\/p><p><span style=\"font-size: 14px;\">La fonction <\/span><span style=\"text-decoration: underline;\"><a href=\"https:\/\/imbalanced-learn.org\/stable\/references\/generated\/imblearn.metrics.classification_report_imbalanced.html\" target=\"_blank\" rel=\"noopener\">classification_report_imbalanced()<\/a><\/span>\u00a0<span style=\"font-size: 14px;\">permet d&rsquo;afficher un rapport contenant notamment les r\u00e9sultats sur l&rsquo;ensemble des m\u00e9triques du package.<\/span><\/p>\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t<section class=\"elementor-section elementor-top-section elementor-element elementor-element-7bb8008 elementor-section-boxed elementor-section-height-default elementor-section-height-default\" data-id=\"7bb8008\" data-element_type=\"section\">\n\t\t\t\t\t\t<div class=\"elementor-container elementor-column-gap-default\">\n\t\t\t\t\t<div class=\"elementor-column elementor-col-100 elementor-top-column elementor-element elementor-element-dc4a08e\" data-id=\"dc4a08e\" data-element_type=\"column\">\n\t\t\t<div class=\"elementor-widget-wrap elementor-element-populated\">\n\t\t\t\t\t\t\t\t<div class=\"elementor-element elementor-element-a67c877 elementor-widget elementor-widget-shortcode\" data-id=\"a67c877\" data-element_type=\"widget\" data-widget_type=\"shortcode.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t\t\t<div class=\"elementor-shortcode\"><script src=\"https:\/\/gist.github.com\/Raphkas15\/a8f007d33300970f9ea62f19eeaa63bf.js\"><\/script><\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-485afc7 elementor-widget elementor-widget-text-editor\" data-id=\"485afc7\" data-element_type=\"widget\" data-widget_type=\"text-editor.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t\t\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t<section class=\"elementor-section elementor-top-section elementor-element elementor-element-1a7c9e4 elementor-section-boxed elementor-section-height-default elementor-section-height-default\" data-id=\"1a7c9e4\" data-element_type=\"section\">\n\t\t\t\t\t\t<div class=\"elementor-container elementor-column-gap-default\">\n\t\t\t\t\t<div class=\"elementor-column elementor-col-100 elementor-top-column elementor-element elementor-element-e711fff\" data-id=\"e711fff\" data-element_type=\"column\">\n\t\t\t<div class=\"elementor-widget-wrap elementor-element-populated\">\n\t\t\t\t\t\t\t\t<div class=\"elementor-element elementor-element-d0dff01 elementor-widget elementor-widget-text-editor\" data-id=\"d0dff01\" data-element_type=\"widget\" data-widget_type=\"text-editor.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t\t\t\t\t<p>Nous obtenons le tableau suivant :\u00a0<\/p>\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t<section class=\"elementor-section elementor-top-section elementor-element elementor-element-62494ec elementor-section-boxed elementor-section-height-default elementor-section-height-default\" data-id=\"62494ec\" data-element_type=\"section\">\n\t\t\t\t\t\t<div class=\"elementor-container elementor-column-gap-default\">\n\t\t\t\t\t<div class=\"elementor-column elementor-col-100 elementor-top-column elementor-element elementor-element-ef89b62\" data-id=\"ef89b62\" data-element_type=\"column\">\n\t\t\t<div class=\"elementor-widget-wrap elementor-element-populated\">\n\t\t\t\t\t\t\t\t<div class=\"elementor-element elementor-element-6f830bd elementor-widget elementor-widget-image\" data-id=\"6f830bd\" data-element_type=\"widget\" data-widget_type=\"image.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<img decoding=\"async\" width=\"512\" height=\"288\" src=\"https:\/\/datascientest.com\/wp-content\/uploads\/2020\/03\/Image2.png\" class=\"attachment-large size-large wp-image-194159\" alt=\"Tableau de classe\" srcset=\"https:\/\/datascientest.com\/wp-content\/uploads\/2020\/03\/Image2.png 512w, https:\/\/datascientest.com\/wp-content\/uploads\/2020\/03\/Image2-300x169.png 300w\" sizes=\"(max-width: 512px) 100vw, 512px\" \/>\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t<section class=\"elementor-section elementor-top-section elementor-element elementor-element-f4032dd elementor-section-boxed elementor-section-height-default elementor-section-height-default\" data-id=\"f4032dd\" data-element_type=\"section\">\n\t\t\t\t\t\t<div class=\"elementor-container elementor-column-gap-default\">\n\t\t\t\t\t<div class=\"elementor-column elementor-col-100 elementor-top-column elementor-element elementor-element-a95ecaa\" data-id=\"a95ecaa\" data-element_type=\"column\">\n\t\t\t<div class=\"elementor-widget-wrap elementor-element-populated\">\n\t\t\t\t\t\t\t\t<div class=\"elementor-element elementor-element-ab12063 elementor-widget elementor-widget-text-editor\" data-id=\"ab12063\" data-element_type=\"widget\" data-widget_type=\"text-editor.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t\t\t\t\t<p>Le tableau montre que le rappel et le f1-score de la classe 1 sont mauvais, tandis que pour la classe 0, ils sont \u00e9lev\u00e9s. En outre, la moyenne g\u00e9om\u00e9trique est \u00e9galement faible.<\/p><p><span style=\"font-size: 14px;\">Le mod\u00e8le entra\u00een\u00e9 n\u2019est donc pas adapt\u00e9 \u00e0 nos donn\u00e9es.<\/span><\/p><p><span style=\"font-size: 14px;\">Dans la\u00a0<\/span><a href=\"https:\/\/datascientest.com\/comment-gerer-les-problemes-de-classification-desequilibree-partie-ii\" target=\"_blank\" rel=\"noopener\"><u>partie II<\/u><\/a><span style=\"font-size: 14px;\"><u>\u00a0<\/u>, nous d\u00e9couvrirons les m\u00e9thodes qui nous permettront d\u2019obtenir de bien meilleurs r\u00e9sultats.<\/span><\/p><p style=\"font-variant-ligatures: normal; font-variant-caps: normal; font-family: 'Proxima Nova', sans-serif; font-size: 14px; font-style: normal; font-weight: 400;\">Envie d\u2019am\u00e9liorer vos comp\u00e9tences pour r\u00e9ussir \u00e0 construire des mod\u00e8les performants et fiables \u00e0 partir d\u2019ensembles de donn\u00e9es d\u00e9s\u00e9quilibr\u00e9s? D\u00e9couvrez tous nos\u00a0<a style=\"font-size: 14px;\" href=\"https:\/\/datascientest.com\/catalogue-formations-data-science\" target=\"_blank\" rel=\"noopener\"><u style=\"font-size: 14px;\">modules d\u2019apprentissage<\/u><\/a>\u00a0!\u00a0<\/p><p style=\"font-variant-ligatures: normal; font-variant-caps: normal; font-family: 'Proxima Nova', sans-serif; font-size: 14px; font-style: normal; font-weight: 400;\">N\u2019h\u00e9sitez pas \u00e0\u00a0<a style=\"font-size: 14px;\" href=\"mailto:contact@datascientest.com\" target=\"_blank\" rel=\"noopener\"><u style=\"font-size: 14px;\">nous contacter<\/u><\/a>\u00a0pour plus d\u2019infos !\u00a0<\/p>\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t\t\t\t<\/div>\n\t\t","protected":false},"excerpt":{"rendered":"<p>Comment g\u00e9rer les probl\u00e8mes de Classification d\u00e9s\u00e9quilibr\u00e9e ? Partie I La classification sur donn\u00e9es d\u00e9s\u00e9quilibr\u00e9es est un probl\u00e8me de classification o\u00f9 l&rsquo;\u00e9chantillon d&rsquo;apprentissage contient une forte disparit\u00e9 entre les classes \u00e0 pr\u00e9dire. Ce probl\u00e8me revient fr\u00e9quemment dans les probl\u00e8mes de classification binaire, et notamment la d\u00e9tection d&rsquo;anomalies. \u00a0Cet article sera divis\u00e9 en deux parties : [&hellip;]<\/p>\n","protected":false},"author":77,"featured_media":184193,"comment_status":"open","ping_status":"open","sticky":false,"template":"elementor_theme","format":"standard","meta":{"content-type":"","inline_featured_image":false},"categories":[2497],"tags":[],"yoast_head":"<!-- This site is optimized with the Yoast SEO Premium plugin v21.7 (Yoast SEO v21.7) - https:\/\/yoast.com\/wordpress\/plugins\/seo\/ -->\n<title>Gestion des proble\u0300mes de Classification de\u0301se\u0301quilibre\u0301e - Partie I<\/title>\n<meta name=\"description\" content=\"La classification sur donn\u00e9es d\u00e9s\u00e9quilibr\u00e9es survient lorsque l\u2019\u00e9chantillon d\u2019apprentissage contient une forte disparit\u00e9 entre les classes \u00e0 pr\u00e9dire.\" \/>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/datascientest.com\/comment-gerer-les-problemes-de-classification-desequilibree-partie-i\" \/>\n<meta property=\"og:locale\" content=\"fr_FR\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"Gestion des proble\u0300mes de Classification de\u0301se\u0301quilibre\u0301e - Partie I\" \/>\n<meta property=\"og:url\" content=\"https:\/\/datascientest.com\/comment-gerer-les-problemes-de-classification-desequilibree-partie-i\" \/>\n<meta property=\"og:site_name\" content=\"Formation Data Science | DataScientest.com\" \/>\n<meta property=\"article:published_time\" content=\"2020-03-25T13:59:54+00:00\" \/>\n<meta property=\"article:modified_time\" content=\"2024-04-23T09:56:50+00:00\" \/>\n<meta property=\"og:image\" content=\"https:\/\/datascientest.com\/wp-content\/uploads\/2022\/12\/reechantillonnage-donnees-desequilibrees.png\" \/>\n\t<meta property=\"og:image:width\" content=\"1200\" \/>\n\t<meta property=\"og:image:height\" content=\"650\" \/>\n\t<meta property=\"og:image:type\" content=\"image\/png\" \/>\n<meta name=\"author\" content=\"Raphael Kassel\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:label1\" content=\"\u00c9crit par\" \/>\n\t<meta name=\"twitter:data1\" content=\"Raphael Kassel\" \/>\n\t<meta name=\"twitter:label2\" content=\"Dur\u00e9e de lecture estim\u00e9e\" \/>\n\t<meta name=\"twitter:data2\" content=\"6 minutes\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\/\/schema.org\",\"@graph\":[{\"@type\":\"WebPage\",\"@id\":\"https:\/\/datascientest.com\/comment-gerer-les-problemes-de-classification-desequilibree-partie-i\",\"url\":\"https:\/\/datascientest.com\/comment-gerer-les-problemes-de-classification-desequilibree-partie-i\",\"name\":\"Gestion des proble\u0300mes de Classification de\u0301se\u0301quilibre\u0301e - Partie I\",\"isPartOf\":{\"@id\":\"https:\/\/datascientest.com\/#website\"},\"datePublished\":\"2020-03-25T13:59:54+00:00\",\"dateModified\":\"2024-04-23T09:56:50+00:00\",\"author\":{\"@id\":\"https:\/\/datascientest.com\/#\/schema\/person\/6eca16f63daaafcbc96e15549b9d4ee1\"},\"description\":\"La classification sur donn\u00e9es d\u00e9s\u00e9quilibr\u00e9es survient lorsque l\u2019\u00e9chantillon d\u2019apprentissage contient une forte disparit\u00e9 entre les classes \u00e0 pr\u00e9dire.\",\"inLanguage\":\"fr-FR\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\/\/datascientest.com\/comment-gerer-les-problemes-de-classification-desequilibree-partie-i\"]}]},{\"@type\":\"WebSite\",\"@id\":\"https:\/\/datascientest.com\/#website\",\"url\":\"https:\/\/datascientest.com\/\",\"name\":\"Formation Data Science | DataScientest.com\",\"description\":\"Leader de la formation en Data Science\",\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\/\/datascientest.com\/?s={search_term_string}\"},\"query-input\":\"required name=search_term_string\"}],\"inLanguage\":\"fr-FR\"},{\"@type\":\"Person\",\"@id\":\"https:\/\/datascientest.com\/#\/schema\/person\/6eca16f63daaafcbc96e15549b9d4ee1\",\"name\":\"Raphael Kassel\",\"image\":{\"@type\":\"ImageObject\",\"inLanguage\":\"fr-FR\",\"@id\":\"https:\/\/datascientest.com\/#\/schema\/person\/image\/\",\"url\":\"https:\/\/datascientest.com\/wp-content\/uploads\/2023\/11\/raph.jpeg\",\"contentUrl\":\"https:\/\/datascientest.com\/wp-content\/uploads\/2023\/11\/raph.jpeg\",\"caption\":\"Raphael Kassel\"},\"description\":\"Raphael Kassel est le Head of Data Programs chez DataScientest. Il supervise une \u00e9quipe Data, impl\u00e9mente de nouveaux parcours comme \\\"Chef de projet IA\\\" et \\\"Data Product Manager\\\", et contribue \u00e0 la modernisation des op\u00e9rations du groupe. En collaboration \u00e9troite avec l'\u00e9quipe de d\u00e9veloppement, il adapte l'accompagnement acad\u00e9mique aux \u00e9volutions de l'\u00e9quipe data et \u00e0 l'internationalisation.\",\"sameAs\":[\"https:\/\/www.linkedin.com\/in\/raphaelkassel\/\",\"Data Science\"],\"url\":\"https:\/\/datascientest.com\/author\/raphaelkassel\"}]}<\/script>\n<!-- \/ Yoast SEO Premium plugin. -->","yoast_head_json":{"title":"Gestion des proble\u0300mes de Classification de\u0301se\u0301quilibre\u0301e - Partie I","description":"La classification sur donn\u00e9es d\u00e9s\u00e9quilibr\u00e9es survient lorsque l\u2019\u00e9chantillon d\u2019apprentissage contient une forte disparit\u00e9 entre les classes \u00e0 pr\u00e9dire.","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:\/\/datascientest.com\/comment-gerer-les-problemes-de-classification-desequilibree-partie-i","og_locale":"fr_FR","og_type":"article","og_title":"Gestion des proble\u0300mes de Classification de\u0301se\u0301quilibre\u0301e - Partie I","og_url":"https:\/\/datascientest.com\/comment-gerer-les-problemes-de-classification-desequilibree-partie-i","og_site_name":"Formation Data Science | DataScientest.com","article_published_time":"2020-03-25T13:59:54+00:00","article_modified_time":"2024-04-23T09:56:50+00:00","og_image":[{"width":1200,"height":650,"url":"https:\/\/datascientest.com\/wp-content\/uploads\/2022\/12\/reechantillonnage-donnees-desequilibrees.png","type":"image\/png"}],"author":"Raphael Kassel","twitter_card":"summary_large_image","twitter_misc":{"\u00c9crit par":"Raphael Kassel","Dur\u00e9e de lecture estim\u00e9e":"6 minutes"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"WebPage","@id":"https:\/\/datascientest.com\/comment-gerer-les-problemes-de-classification-desequilibree-partie-i","url":"https:\/\/datascientest.com\/comment-gerer-les-problemes-de-classification-desequilibree-partie-i","name":"Gestion des proble\u0300mes de Classification de\u0301se\u0301quilibre\u0301e - Partie I","isPartOf":{"@id":"https:\/\/datascientest.com\/#website"},"datePublished":"2020-03-25T13:59:54+00:00","dateModified":"2024-04-23T09:56:50+00:00","author":{"@id":"https:\/\/datascientest.com\/#\/schema\/person\/6eca16f63daaafcbc96e15549b9d4ee1"},"description":"La classification sur donn\u00e9es d\u00e9s\u00e9quilibr\u00e9es survient lorsque l\u2019\u00e9chantillon d\u2019apprentissage contient une forte disparit\u00e9 entre les classes \u00e0 pr\u00e9dire.","inLanguage":"fr-FR","potentialAction":[{"@type":"ReadAction","target":["https:\/\/datascientest.com\/comment-gerer-les-problemes-de-classification-desequilibree-partie-i"]}]},{"@type":"WebSite","@id":"https:\/\/datascientest.com\/#website","url":"https:\/\/datascientest.com\/","name":"Formation Data Science | DataScientest.com","description":"Leader de la formation en Data Science","potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/datascientest.com\/?s={search_term_string}"},"query-input":"required name=search_term_string"}],"inLanguage":"fr-FR"},{"@type":"Person","@id":"https:\/\/datascientest.com\/#\/schema\/person\/6eca16f63daaafcbc96e15549b9d4ee1","name":"Raphael Kassel","image":{"@type":"ImageObject","inLanguage":"fr-FR","@id":"https:\/\/datascientest.com\/#\/schema\/person\/image\/","url":"https:\/\/datascientest.com\/wp-content\/uploads\/2023\/11\/raph.jpeg","contentUrl":"https:\/\/datascientest.com\/wp-content\/uploads\/2023\/11\/raph.jpeg","caption":"Raphael Kassel"},"description":"Raphael Kassel est le Head of Data Programs chez DataScientest. Il supervise une \u00e9quipe Data, impl\u00e9mente de nouveaux parcours comme \"Chef de projet IA\" et \"Data Product Manager\", et contribue \u00e0 la modernisation des op\u00e9rations du groupe. En collaboration \u00e9troite avec l'\u00e9quipe de d\u00e9veloppement, il adapte l'accompagnement acad\u00e9mique aux \u00e9volutions de l'\u00e9quipe data et \u00e0 l'internationalisation.","sameAs":["https:\/\/www.linkedin.com\/in\/raphaelkassel\/","Data Science"],"url":"https:\/\/datascientest.com\/author\/raphaelkassel"}]}},"_links":{"self":[{"href":"https:\/\/datascientest.com\/wp-json\/wp\/v2\/posts\/1203"}],"collection":[{"href":"https:\/\/datascientest.com\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/datascientest.com\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/datascientest.com\/wp-json\/wp\/v2\/users\/77"}],"replies":[{"embeddable":true,"href":"https:\/\/datascientest.com\/wp-json\/wp\/v2\/comments?post=1203"}],"version-history":[{"count":0,"href":"https:\/\/datascientest.com\/wp-json\/wp\/v2\/posts\/1203\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/datascientest.com\/wp-json\/wp\/v2\/media\/184193"}],"wp:attachment":[{"href":"https:\/\/datascientest.com\/wp-json\/wp\/v2\/media?parent=1203"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/datascientest.com\/wp-json\/wp\/v2\/categories?post=1203"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/datascientest.com\/wp-json\/wp\/v2\/tags?post=1203"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}