{"id":1229,"date":"2020-03-25T15:36:57","date_gmt":"2020-03-25T15:36:57","guid":{"rendered":"https:\/\/datascientest.com\/?p=1229"},"modified":"2023-11-09T14:14:59","modified_gmt":"2023-11-09T14:14:59","slug":"comment-gerer-les-problemes-de-classification-desequilibree-partie-ii","status":"publish","type":"post","link":"https:\/\/datascientest.com\/comment-gerer-les-problemes-de-classification-desequilibree-partie-ii","title":{"rendered":"Gestion des proble\u0300mes de Classification de\u0301se\u0301quilibre\u0301e &#8211; Partie II"},"content":{"rendered":"\t\t<div data-elementor-type=\"wp-post\" data-elementor-id=\"1229\" class=\"elementor elementor-1229\" 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-238eb48 elementor-section-boxed elementor-section-height-default elementor-section-height-default\" data-id=\"238eb48\" 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-14561c8\" data-id=\"14561c8\" 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-d0da59f elementor-widget elementor-widget-heading\" data-id=\"d0da59f\" 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 ge\u0301rer les proble\u0300mes de Classification de\u0301se\u0301quilibre\u0301e <br>Partie II<\/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-3c5ccd6 elementor-section-boxed elementor-section-height-default elementor-section-height-default\" data-id=\"3c5ccd6\" 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-f7dcf46\" data-id=\"f7dcf46\" 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-ffd1f41 elementor-widget elementor-widget-text-editor\" data-id=\"ffd1f41\" 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<h2>Les m\u00e9thodes pour lutter contre le d\u00e9s\u00e9quilibre des donn\u00e9es<\/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-e19c71f elementor-section-boxed elementor-section-height-default elementor-section-height-default\" data-id=\"e19c71f\" 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-4576baf\" data-id=\"4576baf\" 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-fb56276 elementor-widget elementor-widget-text-editor\" data-id=\"fb56276\" 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><a href=\"https:\/\/datascientest.com\/comment-gerer-les-problemes-de-classification-desequilibree-partie-i\">Ap<\/a><a href=\"https:\/\/datascientest.com\/comment-gerer-les-problemes-de-classification-desequilibree-partie-i\" target=\"_blank\" rel=\"noopener\">r\u00e8s avoir d\u00e9taill\u00e9 les diff\u00e9rents probl\u00e8mes li\u00e9s au d\u00e9s\u00e9quilibre des donn\u00e9es et d\u00e9montr\u00e9 que le choix de la bonne m\u00e9trique de performance est essentiel pour l\u2019\u00e9valuation de nos mod\u00e8le<\/a>, nous allons pr\u00e9senter une liste non-exhaustive des techniques utiles pour lutter contre ce type de probl\u00e8mes.<\/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-42d7c88 elementor-section-boxed elementor-section-height-default elementor-section-height-default\" data-id=\"42d7c88\" 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-ee75309\" data-id=\"ee75309\" 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-0ab306d elementor-widget elementor-widget-text-editor\" data-id=\"0ab306d\" 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>1) Collecter davantage de donn\u00e9es<\/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-8672c2e elementor-section-boxed elementor-section-height-default elementor-section-height-default\" data-id=\"8672c2e\" 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-019ddb0\" data-id=\"019ddb0\" 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-ed4e223 elementor-widget elementor-widget-text-editor\" data-id=\"ed4e223\" 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>Cela peut para\u00eetre simpliste, mais la collecte de donn\u00e9es suppl\u00e9mentaires est presque toujours n\u00e9glig\u00e9e et peu parfois s\u2019av\u00e9rer efficace.<\/p>\n<p>Pouvez-vous collecter plus de donn\u00e9es ? Prenez quelques minutes pour r\u00e9fl\u00e9chir \u00e0 la possibilit\u00e9 de <b>collecter davantage de donn\u00e9es pour votre probl\u00e8me<\/b>, cela pourrait \u00e9ventuellement <b>r\u00e9\u00e9quilibrer vos classes <\/b>\u00e0 un degr\u00e9 variable.<\/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-e37dbfb elementor-section-boxed elementor-section-height-default elementor-section-height-default\" data-id=\"e37dbfb\" 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-b40ea05\" data-id=\"b40ea05\" 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-91c3c8e elementor-widget elementor-widget-text-editor\" data-id=\"91c3c8e\" 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>2) Utiliser des m\u00e9thodes de r\u00e9\u00e9chantillonnage\u00a0<\/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-f0f2ecf elementor-section-boxed elementor-section-height-default elementor-section-height-default\" data-id=\"f0f2ecf\" 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-ec1271c\" data-id=\"ec1271c\" 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-f33cb9e elementor-widget elementor-widget-text-editor\" data-id=\"f33cb9e\" 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>Vous pouvez modifier l&rsquo;ensemble de donn\u00e9es que vous utilisez avant d\u2019entra\u00eener votre mod\u00e8le pr\u00e9dictif afin d&rsquo;avoir des donn\u00e9es plus \u00e9quilibr\u00e9es.<\/p><p>Cette strat\u00e9gie est appel\u00e9e r\u00e9\u00e9chantillonnage et il existe deux m\u00e9thodes principales que vous pouvez utiliser pour \u00e9galiser les classes :\u00a0<\/p><p>Le <b>sur-\u00e9chantillonnage<\/b> : Oversampling et le <b>sous-\u00e9chantillonnage<\/b> :Undersampling.\u00a0<\/p><p>Les m\u00e9thodes d&rsquo;<b>Oversampling<\/b> fonctionnent en augmentant le nombre d&rsquo;observations de la (des) classe(s) minoritaire(s) afin d&rsquo;arriver \u00e0 un ratio classe minoritaire\/ classe majoritaire satisfaisant.<\/p><p>\u00a0<span style=\"font-size: 14px;\">Les m\u00e9thodes d&rsquo;<b>Undersampling<\/b> fonctionnent en diminuant le nombre d&rsquo;observations de la\u00a0 (des) classe(s) majoritaire(s) afin d&rsquo;arriver \u00e0 un ratio classe minoritaire\/ classe majoritaire satisfaisant.\u00a0<\/span><\/p><p><span style=\"font-size: 14px;\">Sur Python, le package imblearn proposent un panel de m\u00e9thodes de r\u00e9\u00e9chantillonnage disponibles\u00a0<\/span><a href=\"https:\/\/pypi.org\/project\/imbalanced-learn\/\" target=\"_blank\" rel=\"noopener\"><u>ici<\/u><\/a><span style=\"font-size: 14px;\">.\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-728ea19 elementor-section-boxed elementor-section-height-default elementor-section-height-default\" data-id=\"728ea19\" 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-7a63e72\" data-id=\"7a63e72\" 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-b4ac3fd elementor-widget elementor-widget-shortcode\" data-id=\"b4ac3fd\" 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\/82886a71165f63dfff17edc1df8c22c1.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-8439014 elementor-section-boxed elementor-section-height-default elementor-section-height-default\" data-id=\"8439014\" 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-34dfa92\" data-id=\"34dfa92\" 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-f16f8b7 elementor-widget elementor-widget-text-editor\" data-id=\"f16f8b7\" 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>Ces approches sont tr\u00e8s faciles \u00e0 mettre en \u0153uvre et rapides \u00e0 ex\u00e9cuter. Elles constituent un excellent point de d\u00e9part.<\/p><blockquote><span style=\"font-size: 14px;\">Notre conseil: toujours essayer les deux approches sur tous vos ensembles de donn\u00e9es d\u00e9s\u00e9quilibr\u00e9es,\u00a0 et v\u00e9rifier si cela am\u00e9liore vos m\u00e9triques de performances choisies.<\/span><\/blockquote><ul><li>Privil\u00e9giez le sous-\u00e9chantillonnage lorsque vous disposez de grands jeux de donn\u00e9es : des dizaines ou des centaines de milliers de cas ou plus.<\/li><li>Privil\u00e9giez le sur-\u00e9chantillonnage lorsque vous ne disposez pas de beaucoup de donn\u00e9es :des dizaines de milliers ou moins.<\/li><\/ul><p>Envisagez de tester diff\u00e9rents ratios de classes. Par exemple, vous ne devez pas obligatoirement viser un ratio 1:1 dans un probl\u00e8me de classification binaire, essayez d&rsquo;autres ratios.<\/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-6b4b312 elementor-section-boxed elementor-section-height-default elementor-section-height-default\" data-id=\"6b4b312\" 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-f63d1ee\" data-id=\"f63d1ee\" 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-2707332 elementor-widget elementor-widget-text-editor\" data-id=\"2707332\" 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>3) La g\u00e9n\u00e9ration d\u2019\u00e9chantillons synth\u00e9tiques<\/h3>\n<div>\n<div>Il existe des algorithmes pour <b>g\u00e9n\u00e9rer des \u00e9chantillons synth\u00e9tiques de mani\u00e8re automatique<\/b>. Le plus populaire de ces algorithmes est <b>SMOTE <\/b>(pour Synthetic Minority Over-sampling Technique). Comme son nom l&rsquo;indique, SMOTE est une <b>m\u00e9thode de sur-\u00e9chantillonnage<\/b>. Elle fonctionne en cr\u00e9ant des \u00e9chantillons synth\u00e9tiques \u00e0 partir de la classe minoritaire au lieu de cr\u00e9er de simples copies. Pour en savoir plus sur SMOTE, consultez l&rsquo;<a href=\"https:\/\/imbalanced-learn.org\/stable\/references\/generated\/imblearn.over_sampling.SMOTE.html\" target=\"_blank\">article original<\/a><u>.<\/u><\/div>\n<div>L\u2019algorithme&nbsp;<a href=\"https:\/\/datascientest.com\/algorithme-des-k-means\" target=\"_blank\" rel=\"noopener\"><u>ClusterCentroids<\/u><\/a>&nbsp;est un <b>algorithme d\u2019Undersampling<\/b>&nbsp; qui utilise des m\u00e9thodes de <b>Clustering<\/b> pour g\u00e9n\u00e9rer un certain nombre de centro\u00efdes \u00e0 partir des donn\u00e9es d&rsquo;origine, afin de perdre le moins d\u2019information possible sur la classe majoritaire, lorsque celle-ci doit \u00eatre r\u00e9duite.<\/div>\n<\/div>\n<div>&nbsp;<\/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-8326c9b elementor-section-boxed elementor-section-height-default elementor-section-height-default\" data-id=\"8326c9b\" 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-ca34743\" data-id=\"ca34743\" 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-92ebe5c elementor-widget elementor-widget-shortcode\" data-id=\"92ebe5c\" 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\/4d8f9f412777b591abb20629f45d4d0a.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-72796f3 elementor-section-boxed elementor-section-height-default elementor-section-height-default\" data-id=\"72796f3\" 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-44f2db3\" data-id=\"44f2db3\" 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-db552b0 elementor-widget elementor-widget-text-editor\" data-id=\"db552b0\" 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>4) Repenser le probl\u00e8me<\/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-3eaee94 elementor-section-boxed elementor-section-height-default elementor-section-height-default\" data-id=\"3eaee94\" 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-04abed0\" data-id=\"04abed0\" 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-238aa35 elementor-widget elementor-widget-text-editor\" data-id=\"238aa35\" 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: 16px; color: rgb(0, 0, 0);\">Parfois, les m\u00e9thodes de r\u00e9\u00e9chantillonnage ne sont pas assez efficaces, et dans ce cas il convient de repenser le probl\u00e8me. Il peut s\u2019agir de l\u2019algorithme utilis\u00e9 qui ne convient pas \u00e0 vos donn\u00e9es.&nbsp;<\/span><\/h3><h3><span style=\"font-size: 16px; color: rgb(0, 0, 0); font-weight: normal;\">N\u2019h\u00e9sitez pas \u00e0 en tester d\u2019autres, combin\u00e9s \u00e9ventuellement aux m\u00e9thodes de r\u00e9\u00e9chantillonnage vues ci-dessus.<\/span><br><\/h3>\n<p>Les mod\u00e8les d\u2019ensemble \u00e0 base d\u2019arbres comme les RandomForest sont g\u00e9n\u00e9ralement plus adapt\u00e9s aux donn\u00e9es d\u00e9s\u00e9quilibr\u00e9es.<\/p>\n<p>Il est \u00e9galement possible de <b>jouer avec les probabilit\u00e9s<\/b>. Si l&rsquo;on souhaite par exemple, pouvoir absolument pr\u00e9dire la grande majorit\u00e9 des churners potentiels, quitte \u00e0 mal classer quelques non-churners, on peut modifier le seuil de probabilit\u00e9 au del\u00e0 duquel les clients sont consid\u00e9r\u00e9s comme churners.<\/p>\n<p>Plus le seuil est bas, plus la pr\u00e9cision de notre classe augmentera, mais le rappel diminuera.<\/p>\n<div>&nbsp;<\/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-ed4e808 elementor-section-boxed elementor-section-height-default elementor-section-height-default\" data-id=\"ed4e808\" 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-bf318bb\" data-id=\"bf318bb\" 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-ae3bbfa elementor-widget elementor-widget-shortcode\" data-id=\"ae3bbfa\" 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\/910ed14bd6e2e5ec83502f00d3000075.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-f861f3f elementor-section-boxed elementor-section-height-default elementor-section-height-default\" data-id=\"f861f3f\" 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-2df167b\" data-id=\"2df167b\" 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-b945eff elementor-widget elementor-widget-text-editor\" data-id=\"b945eff\" 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>5) Utiliser un mod\u00e8le p\u00e9nalis\u00e9<\/h3>\n<p>La <b>classification p\u00e9nalis\u00e9e<\/b> permet d\u2019imposer un co\u00fbt suppl\u00e9mentaire au mod\u00e8le pour les erreurs de classification commises sur la classe minoritaire pendant la formation. Ces p\u00e9nalit\u00e9s peuvent biaiser le mod\u00e8le pour qu&rsquo;il accorde plus d&rsquo;attention \u00e0 la classe minoritaire.<\/p>\n<p>Dans la plupart des <b>classes d&rsquo;algorithmes de scikit-learn<\/b>, il est possible de simplement utiliser le param\u00e8tre `class_weight`. Il permet de p\u00e9naliser les erreurs faites sur une classe par un nouveau poids.&nbsp;<\/p><blockquote>Plus le poids d&rsquo;une classe est important, plus les erreurs sur cette classe sont p\u00e9nalis\u00e9es, et plus on lui donne de l&rsquo;importance.<\/blockquote>\n<p>Les poids doivent \u00eatre indiqu\u00e9s sous forme de dictionnaire, par exemple : `{0:1, 1:5}`, pour donner 5 fois plus de poids aux erreurs faites sur la classe 1.&nbsp;<\/p>\n<p>L&rsquo;argument `\u00a0\u00bbbalanced\u00a0\u00bb` permet d&rsquo;associer \u00e0 chaque classe un poids inversement proportionnel \u00e0 sa fr\u00e9quence.<\/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-974d034 elementor-section-boxed elementor-section-height-default elementor-section-height-default\" data-id=\"974d034\" 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-d17fe60\" data-id=\"d17fe60\" 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-5a29421 elementor-widget elementor-widget-shortcode\" data-id=\"5a29421\" 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\/982e8e5d95945fe57aea303357ba9f74.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-b4bad55 elementor-section-boxed elementor-section-height-default elementor-section-height-default\" data-id=\"b4bad55\" 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-fb66771\" data-id=\"fb66771\" 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-3856064 elementor-widget elementor-widget-text-editor\" data-id=\"3856064\" 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>6) Utiliser des m\u00e9thodes g\u00e9n\u00e9rant des sous-ensembles sous-\u00e9chantillonn\u00e9s&nbsp;<\/h3>\n<p>Une autre solution propos\u00e9e par le module&nbsp; imblearn.ensemble, est l&rsquo;utilisation de classes contenant des <b>mod\u00e8les d&rsquo;ensembles comme le Boosting ou Bagging <\/b>qui sont entra\u00een\u00e9s \u00e0 chaque \u00e9tape de l&rsquo;algorithme sur un \u00e9chantillon r\u00e9\u00e9quilibr\u00e9 automatiquement entre les diff\u00e9rentes classes.<\/p>\n<p><span style=\"font-size: 14px;\">Ces impl\u00e9mentations de mod\u00e8les permettent de se passer de m\u00e9thodes de r\u00e9\u00e9chantillonnage avant l&rsquo;entra\u00eenement, et de les appliquer de mani\u00e8re automatique \u00e0 chaque s\u00e9lection de donn\u00e9es par l&rsquo;algorithme.<\/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-1415844 elementor-section-boxed elementor-section-height-default elementor-section-height-default\" data-id=\"1415844\" 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-54990a7\" data-id=\"54990a7\" 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-482f0fb elementor-widget elementor-widget-shortcode\" data-id=\"482f0fb\" 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\/84db253c0e288f5622750d397f74af49.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-401fbab elementor-section-boxed elementor-section-height-default elementor-section-height-default\" data-id=\"401fbab\" 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-05a61b3\" data-id=\"05a61b3\" 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-b3780fd elementor-widget elementor-widget-text-editor\" data-id=\"b3780fd\" 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>7) Soyez cr\u00e9atifs&nbsp;<\/h3>\n<p>Vous pouvez <b>tester l\u2019ensemble de ces techniques, les combiner, ou m\u00eame penser \u00e0 relabeliser les donn\u00e9es de la classe majoritaire en sous-classes<\/b> pour obtenir un probl\u00e8me de classification multi-classes plus \u00e9quilibr\u00e9.<\/p>\n<p>Dans certains cas, vous pouvez \u00e9galement penser \u00e0 utiliser d\u2019autre m\u00e9thodes de Machine Learning comme l\u2019 Anomaly Detection ou l\u2019 Active Learning.<\/p>\n<p>Nous avons pr\u00e9sent\u00e9 un certain nombre de techniques, parmi lesquelles vous pourrez choisir lorsque vous serez amen\u00e9s \u00e0 travailler avec ce genre de donn\u00e9es. N\u2019h\u00e9sitez pas \u00e0 tester ces m\u00e9thodes individuellement, et commencez par les plus simples !<\/p>\n<p>Envie d\u2019am\u00e9liorer vos comp\u00e9tences pour r\u00e9ussir \u00e0 construire des mod\u00e8les performants et fiables \u00e0 partir d&rsquo;ensembles de donn\u00e9es d\u00e9s\u00e9quilibr\u00e9s? D\u00e9couvrez tous nos&nbsp;<a href=\"https:\/\/datascientest.com\/catalogue-formations-data-science\" target=\"_blank\" rel=\"noopener\"><u>modules d\u2019apprentissage<\/u><\/a>&nbsp;!&nbsp;<\/p>\n<p>N&rsquo;h\u00e9sitez pas \u00e0&nbsp;<a href=\"mailto:contact@datascientest.com\" target=\"_blank\" rel=\"noopener\"><u>nous contacter<\/u><\/a>&nbsp;pour plus d&rsquo;infos !&nbsp;<\/p>\n<div>&nbsp;<\/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\t\t\t<\/div>\n\t\t","protected":false},"excerpt":{"rendered":"<p>Comment ge\u0301rer les proble\u0300mes de Classification de\u0301se\u0301quilibre\u0301e Partie II Les m\u00e9thodes pour lutter contre le d\u00e9s\u00e9quilibre des donn\u00e9es Apr\u00e8s avoir d\u00e9taill\u00e9 les diff\u00e9rents probl\u00e8mes li\u00e9s au d\u00e9s\u00e9quilibre des donn\u00e9es et d\u00e9montr\u00e9 que le choix de la bonne m\u00e9trique de performance est essentiel pour l\u2019\u00e9valuation de nos mod\u00e8le, nous allons pr\u00e9senter une liste non-exhaustive des techniques [&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 II<\/title>\n<meta name=\"description\" content=\"Comment g\u00e9rer les probl\u00e8mes de Classification d\u00e9s\u00e9quilibr\u00e9e ? D\u00e9couvrez les m\u00e9thodes pour lutter contre le d\u00e9s\u00e9quilibre des donn\u00e9es !\" \/>\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-ii\" \/>\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 II\" \/>\n<meta property=\"og:url\" content=\"https:\/\/datascientest.com\/comment-gerer-les-problemes-de-classification-desequilibree-partie-ii\" \/>\n<meta property=\"og:site_name\" content=\"Formation Data Science | DataScientest.com\" \/>\n<meta property=\"article:published_time\" content=\"2020-03-25T15:36:57+00:00\" \/>\n<meta property=\"article:modified_time\" content=\"2023-11-09T14:14:59+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=\"5 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-ii\",\"url\":\"https:\/\/datascientest.com\/comment-gerer-les-problemes-de-classification-desequilibree-partie-ii\",\"name\":\"Gestion des proble\u0300mes de Classification de\u0301se\u0301quilibre\u0301e - Partie II\",\"isPartOf\":{\"@id\":\"https:\/\/datascientest.com\/#website\"},\"datePublished\":\"2020-03-25T15:36:57+00:00\",\"dateModified\":\"2023-11-09T14:14:59+00:00\",\"author\":{\"@id\":\"https:\/\/datascientest.com\/#\/schema\/person\/6eca16f63daaafcbc96e15549b9d4ee1\"},\"description\":\"Comment g\u00e9rer les probl\u00e8mes de Classification d\u00e9s\u00e9quilibr\u00e9e ? D\u00e9couvrez les m\u00e9thodes pour lutter contre le d\u00e9s\u00e9quilibre des donn\u00e9es !\",\"inLanguage\":\"fr-FR\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\/\/datascientest.com\/comment-gerer-les-problemes-de-classification-desequilibree-partie-ii\"]}]},{\"@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 II","description":"Comment g\u00e9rer les probl\u00e8mes de Classification d\u00e9s\u00e9quilibr\u00e9e ? D\u00e9couvrez les m\u00e9thodes pour lutter contre le d\u00e9s\u00e9quilibre des donn\u00e9es !","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-ii","og_locale":"fr_FR","og_type":"article","og_title":"Gestion des proble\u0300mes de Classification de\u0301se\u0301quilibre\u0301e - Partie II","og_url":"https:\/\/datascientest.com\/comment-gerer-les-problemes-de-classification-desequilibree-partie-ii","og_site_name":"Formation Data Science | DataScientest.com","article_published_time":"2020-03-25T15:36:57+00:00","article_modified_time":"2023-11-09T14:14:59+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":"5 minutes"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"WebPage","@id":"https:\/\/datascientest.com\/comment-gerer-les-problemes-de-classification-desequilibree-partie-ii","url":"https:\/\/datascientest.com\/comment-gerer-les-problemes-de-classification-desequilibree-partie-ii","name":"Gestion des proble\u0300mes de Classification de\u0301se\u0301quilibre\u0301e - Partie II","isPartOf":{"@id":"https:\/\/datascientest.com\/#website"},"datePublished":"2020-03-25T15:36:57+00:00","dateModified":"2023-11-09T14:14:59+00:00","author":{"@id":"https:\/\/datascientest.com\/#\/schema\/person\/6eca16f63daaafcbc96e15549b9d4ee1"},"description":"Comment g\u00e9rer les probl\u00e8mes de Classification d\u00e9s\u00e9quilibr\u00e9e ? D\u00e9couvrez les m\u00e9thodes pour lutter contre le d\u00e9s\u00e9quilibre des donn\u00e9es !","inLanguage":"fr-FR","potentialAction":[{"@type":"ReadAction","target":["https:\/\/datascientest.com\/comment-gerer-les-problemes-de-classification-desequilibree-partie-ii"]}]},{"@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\/1229"}],"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=1229"}],"version-history":[{"count":0,"href":"https:\/\/datascientest.com\/wp-json\/wp\/v2\/posts\/1229\/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=1229"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/datascientest.com\/wp-json\/wp\/v2\/categories?post=1229"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/datascientest.com\/wp-json\/wp\/v2\/tags?post=1229"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}