{"id":9702,"date":"2020-09-18T08:47:38","date_gmt":"2020-09-18T07:47:38","guid":{"rendered":"https:\/\/datascientest.com\/?p=9702"},"modified":"2023-11-09T14:12:17","modified_gmt":"2023-11-09T14:12:17","slug":"nlp-word-embedding-word2vec","status":"publish","type":"post","link":"https:\/\/datascientest.com\/nlp-word-embedding-word2vec","title":{"rendered":"Word2vec : NLP &amp; Word Embedding"},"content":{"rendered":"\t\t<div data-elementor-type=\"wp-post\" data-elementor-id=\"9702\" class=\"elementor elementor-9702\" data-elementor-post-type=\"post\">\n\t\t\t\t\t\t\t<div class=\"elementor-element elementor-element-e1d280a e-flex e-con-boxed e-con\" data-id=\"e1d280a\" data-element_type=\"container\" data-settings=\"{&quot;content_width&quot;:&quot;boxed&quot;}\">\n\t\t\t\t\t<div class=\"e-con-inner\">\n\t\t\t\t<div class=\"elementor-element elementor-element-a0dabe8 elementor-widget elementor-widget-heading\" data-id=\"a0dabe8\" 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\">Le word embedding d\u00e9signe un ensemble de m\u00e9thode d\u2019apprentissage visant \u00e0 repr\u00e9senter les mots d\u2019un texte par des vecteurs de nombres r\u00e9els.\u00a0 Aujourd\u2019hui, nous allons vous pr\u00e9senter le 3\u00e8me volet de notre dossier NLP.\u00a0<\/h2>\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t<div class=\"elementor-element elementor-element-811b324 e-flex e-con-boxed e-con\" data-id=\"811b324\" data-element_type=\"container\" data-settings=\"{&quot;content_width&quot;:&quot;boxed&quot;}\">\n\t\t\t\t\t<div class=\"e-con-inner\">\n\t\t\t\t<div class=\"elementor-element elementor-element-e30642c elementor-widget elementor-widget-text-editor\" data-id=\"e30642c\" 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><b>Vous avez rat\u00e9 les premiers \u00e9pisodes ? Pas de panique les voici :\u00a0<\/b><\/p><ol><li><a href=\"https:\/\/datascientest.com\/introduction-au-nlp-natural-language-processing\">Introduction au NLP<\/a><\/li><li><a href=\"https:\/\/datascientest.com\/nlp-word-embedding-word2vec\">Word embedding &#8211; Word2vec<\/a><ol><li><a href=\"#wordembedding\">D\u00e9finition word embedding<\/a><\/li><li><a href=\"#word2vec\">Embedding \u00e0 l&rsquo;aide du Word2vec<\/a><\/li><li><a href=\"#metrique\">M\u00e9trique dans cette espace<\/a><\/li><\/ol><\/li><li><a href=\"https:\/\/datascientest.com\/nlp-word-translation\">NLP &#8211; Word translation<\/a><\/li><li><a href=\"https:\/\/datascientest.com\/nlp-twitter-analyse-de-sentiment\">NLP Twitter &#8211; Analyse des sentiments<\/a><\/li><\/ol>\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-6087383 elementor-widget elementor-widget-text-editor\" data-id=\"6087383\" 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-weight: 400;\">Cette section a pour but d\u2019expliquer le fonctionnement et d\u2019impl\u00e9menter le c\u00e9l\u00e8bre <\/span><b>algorithme Word2vec sur Python<\/b><span style=\"font-weight: 400;\">.\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\t\t<\/div>\n\t\t<div class=\"elementor-element elementor-element-12b7873 e-flex e-con-boxed e-con\" data-id=\"12b7873\" data-element_type=\"container\" data-settings=\"{&quot;content_width&quot;:&quot;boxed&quot;}\">\n\t\t\t\t\t<div class=\"e-con-inner\">\n\t\t\t\t<div class=\"elementor-element elementor-element-ddef68f elementor-widget elementor-widget-menu-anchor\" data-id=\"ddef68f\" data-element_type=\"widget\" data-widget_type=\"menu-anchor.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t<style>\/*! elementor - v3.15.0 - 02-08-2023 *\/\nbody.elementor-page .elementor-widget-menu-anchor{margin-bottom:0}<\/style>\t\t<div id=\"wordembedding\" class=\"elementor-menu-anchor\"><\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-50eca36 elementor-widget elementor-widget-heading\" data-id=\"50eca36\" data-element_type=\"widget\" data-widget_type=\"heading.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t<h3 class=\"elementor-heading-title elementor-size-default\">Word Embedding<\/h3>\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-acdb110 elementor-widget elementor-widget-text-editor\" data-id=\"acdb110\" 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<blockquote><p><span style=\"font-weight: 400;\">Pour rappel, le word embedding est capable <\/span><b>en\u00a0r\u00e9duisant la dimension<\/b><b>de\u00a0capturer le contexte, la similarit\u00e9 s\u00e9mantique et syntaxique (genre, synonymes, &#8230;) d&rsquo;un mot<\/b><span style=\"font-weight: 400;\">. <\/span><\/p><\/blockquote><p><span style=\"font-weight: 400;\">Par exemple, on pourrait s&rsquo;attendre \u00e0 ce que les mots \u00ab remarquable \u00bb et \u00ab admirable \u00bb soient repr\u00e9sent\u00e9s par des vecteurs relativement peu distants dans l&rsquo;espace vectoriel o\u00f9 sont d\u00e9finis ces vecteurs.<\/span><\/p><p><span style=\"font-weight: 400;\">La m\u00e9thode d\u2019embedding g\u00e9n\u00e9ralement utilis\u00e9e pour <\/span><b>r\u00e9duire la dimension d\u2019un vecteur<\/b><span style=\"font-weight: 400;\"> consiste \u00e0 utiliser le r\u00e9sultat que retourne une couche dense, c&rsquo;est \u00e0 dire de multiplier une matrice d\u2019embedding\u00a0W\u00a0par la repr\u00e9sentation \u00ab\u00a0one hot\u00a0\u00bb du mot :<\/span><\/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-256f9a6 elementor-widget elementor-widget-image\" data-id=\"256f9a6\" 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=\"260\" height=\"144\" src=\"https:\/\/datascientest.com\/wp-content\/uploads\/2020\/09\/Capture-de\u0301cran-2020-09-18-a\u0300-09.23.47.png\" class=\"attachment-full size-full wp-image-9721\" alt=\"Word2vec\" \/>\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<div class=\"elementor-element elementor-element-fd924f8 elementor-widget elementor-widget-text-editor\" data-id=\"fd924f8\" 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>Sous forme vectorielle :<\/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-548c2cd elementor-widget elementor-widget-image\" data-id=\"548c2cd\" 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=\"278\" height=\"78\" src=\"https:\/\/datascientest.com\/wp-content\/uploads\/2020\/09\/Capture-de\u0301cran-2020-09-18-a\u0300-09.23.58.png\" class=\"attachment-full size-full wp-image-9719\" alt=\"Word2vec\" \/>\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<div class=\"elementor-element elementor-element-3859302 elementor-widget elementor-widget-image\" data-id=\"3859302\" 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<figure class=\"wp-caption\">\n\t\t\t\t\t\t\t\t\t\t\t<a href=\"https:\/\/medium.com\/datadriveninvestor\/deep-learning-autoencoders-db265359943e\">\n\t\t\t\t\t\t\t<img decoding=\"async\" width=\"2560\" height=\"2675\" src=\"https:\/\/datascientest.com\/wp-content\/uploads\/2020\/09\/Sche\u0301ma-1.png\" class=\"attachment-full size-full wp-image-9809\" alt=\"auto encodeur\" srcset=\"https:\/\/datascientest.com\/wp-content\/uploads\/2020\/09\/Sche\u0301ma-1.png 2560w, https:\/\/datascientest.com\/wp-content\/uploads\/2020\/09\/Sche\u0301ma-1-287x300.png 287w, https:\/\/datascientest.com\/wp-content\/uploads\/2020\/09\/Sche\u0301ma-1-980x1024.png 980w, https:\/\/datascientest.com\/wp-content\/uploads\/2020\/09\/Sche\u0301ma-1-768x802.png 768w, https:\/\/datascientest.com\/wp-content\/uploads\/2020\/09\/Sche\u0301ma-1-1470x1536.png 1470w, https:\/\/datascientest.com\/wp-content\/uploads\/2020\/09\/Sche\u0301ma-1-1960x2048.png 1960w\" sizes=\"(max-width: 2560px) 100vw, 2560px\" \/>\t\t\t\t\t\t\t\t<\/a>\n\t\t\t\t\t\t\t\t\t\t\t<figcaption class=\"widget-image-caption wp-caption-text\">Auto-encodeur<\/figcaption>\n\t\t\t\t\t\t\t\t\t\t<\/figure>\n\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\t\t<\/div>\n\t\t<div class=\"elementor-element elementor-element-d64c033 e-flex e-con-boxed e-con\" data-id=\"d64c033\" data-element_type=\"container\" data-settings=\"{&quot;content_width&quot;:&quot;boxed&quot;}\">\n\t\t\t\t\t<div class=\"e-con-inner\">\n\t\t\t\t<div class=\"elementor-element elementor-element-d5e0e6a elementor-widget elementor-widget-menu-anchor\" data-id=\"d5e0e6a\" data-element_type=\"widget\" data-widget_type=\"menu-anchor.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t\t\t<div id=\"word2vec\" class=\"elementor-menu-anchor\"><\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-4e771b4 elementor-widget elementor-widget-heading\" data-id=\"4e771b4\" data-element_type=\"widget\" data-widget_type=\"heading.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t<h3 class=\"elementor-heading-title elementor-size-default\">Embedding \u00e0 l\u2019aide du Word2vec<\/h3>\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-a238c02 elementor-widget elementor-widget-text-editor\" data-id=\"a238c02\" 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-weight: 400;\">En revanche, pour \u00e9viter les contraintes de la premi\u00e8re m\u00e9thode, il est possible d&rsquo;entra\u00eener la matrice\u00a0W\u00a0de <\/span><b>mani\u00e8re non supervis\u00e9<\/b><span style=\"font-weight: 400;\"> avec <\/span><b>simplement du texte<\/b><span style=\"font-weight: 400;\"> \u00e0 l\u2019aide du c\u00e9l\u00e8bre <\/span><b>algorithme Word2Vec.<\/b><\/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-2941354 elementor-widget elementor-widget-heading\" data-id=\"2941354\" data-element_type=\"widget\" data-widget_type=\"heading.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t<h4 class=\"elementor-heading-title elementor-size-medium\">Comment word2vec fonctionne ?<\/h4>\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-07e8959 elementor-widget elementor-widget-text-editor\" data-id=\"07e8959\" 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<blockquote><p><span style=\"font-weight: 400;\">Il existe deux variantes du Word2vec, les deux utilisent un r\u00e9seau de neurones \u00e0 3 couches (1 couche d&rsquo;entr\u00e9e, 1 couche cach\u00e9e, 1 couche de sortie) : Common Bag Of Words (CBOW) et Skip-gram.<\/span><\/p><\/blockquote><p><span style=\"font-weight: 400;\">Dans l&rsquo;image suivante, le <\/span><b>mot dans la case bleu est appel\u00e9 le\u00a0mot cible<\/b><span style=\"font-weight: 400;\">\u00a0et <\/span><b>les mots dans les cases blanches sont appel\u00e9s\u00a0mots du contexte<\/b><span style=\"font-weight: 400;\">\u00a0dans <\/span><b>une\u00a0fen\u00eatre\u00a0de taille 5<\/b><span style=\"font-weight: 400;\">.<\/span><\/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-3c047d4 elementor-widget elementor-widget-image\" data-id=\"3c047d4\" 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=\"842\" height=\"334\" src=\"https:\/\/datascientest.com\/wp-content\/uploads\/2020\/09\/Capture-de\u0301cran-2020-09-18-a\u0300-09.27.50.png\" class=\"attachment-full size-full wp-image-9722\" alt=\"Embedding \u00e0 l\u2019aide du Word2vec\" srcset=\"https:\/\/datascientest.com\/wp-content\/uploads\/2020\/09\/Capture-de\u0301cran-2020-09-18-a\u0300-09.27.50.png 842w, https:\/\/datascientest.com\/wp-content\/uploads\/2020\/09\/Capture-de\u0301cran-2020-09-18-a\u0300-09.27.50-300x119.png 300w, https:\/\/datascientest.com\/wp-content\/uploads\/2020\/09\/Capture-de\u0301cran-2020-09-18-a\u0300-09.27.50-768x305.png 768w\" sizes=\"(max-width: 842px) 100vw, 842px\" \/>\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<div class=\"elementor-element elementor-element-c91079d elementor-widget elementor-widget-text-editor\" data-id=\"c91079d\" 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><b>CBOW\u00a0:<\/b> <b>Le mod\u00e8le est nourri par le contexte<\/b><span style=\"font-weight: 400;\">, et pr\u00e9dit le mot cible. Le r\u00e9sultat de la couche cach\u00e9e est la nouvelle repr\u00e9sentation du mot (<\/span><span style=\"font-weight: 400;\">\u210e<\/span><span style=\"font-weight: 400;\">1, &#8230;, <\/span><span style=\"font-weight: 400;\">\u210e\ud835\udc41<\/span><span style=\"font-weight: 400;\">).<\/span><\/p><p><b>Skip Gram\u00a0:<\/b> <b>Le mod\u00e8le est nourri par le mot cible<\/b><span style=\"font-weight: 400;\">, et pr\u00e9dit les mots du contexte. Le r\u00e9sultat de la couche cach\u00e9e est la nouvelle repr\u00e9sentation du mot (<\/span><span style=\"font-weight: 400;\">\u210e<\/span><span style=\"font-weight: 400;\">1, &#8230;, <\/span><span style=\"font-weight: 400;\">\u210e\ud835\udc41<\/span><span style=\"font-weight: 400;\">).<\/span><\/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-20f8e31 elementor-widget elementor-widget-image\" data-id=\"20f8e31\" 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=\"720\" height=\"433\" src=\"https:\/\/datascientest.com\/wp-content\/uploads\/2020\/09\/word-2-vec-illu.png\" class=\"attachment-full size-full wp-image-9810\" alt=\"word2vec\" srcset=\"https:\/\/datascientest.com\/wp-content\/uploads\/2020\/09\/word-2-vec-illu.png 720w, https:\/\/datascientest.com\/wp-content\/uploads\/2020\/09\/word-2-vec-illu-300x180.png 300w\" sizes=\"(max-width: 720px) 100vw, 720px\" \/>\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<div class=\"elementor-element elementor-element-5678dfe elementor-widget elementor-widget-heading\" data-id=\"5678dfe\" data-element_type=\"widget\" data-widget_type=\"heading.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t<h4 class=\"elementor-heading-title elementor-size-medium\">Mise en forme des donn\u00e9es<\/h4>\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-8d8bc35 elementor-widget elementor-widget-text-editor\" data-id=\"8d8bc35\" 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-weight: 400;\">Ici, nous allons <\/span><b>pr\u00e9senter le mod\u00e8le\u00a0CBOW<\/b><span style=\"font-weight: 400;\">\u00a0c&rsquo;est-\u00e0-dire que le <\/span><b>contexte est l&rsquo;entr\u00e9e<\/b><span style=\"font-weight: 400;\"> de notre mod\u00e8le et le <\/span><b>mot cible (mot bleu) est la sortie<\/b><span style=\"font-weight: 400;\">. Nous d\u00e9finissons une fen\u00eatre de longueurs 5 pour le contexte (entr\u00e9e).<\/span><\/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-0f9e351 elementor-widget elementor-widget-image\" data-id=\"0f9e351\" 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=\"842\" height=\"334\" src=\"https:\/\/datascientest.com\/wp-content\/uploads\/2020\/09\/Capture-de\u0301cran-2020-09-18-a\u0300-09.27.50.png\" class=\"attachment-full size-full wp-image-9722\" alt=\"Embedding \u00e0 l\u2019aide du Word2vec\" srcset=\"https:\/\/datascientest.com\/wp-content\/uploads\/2020\/09\/Capture-de\u0301cran-2020-09-18-a\u0300-09.27.50.png 842w, https:\/\/datascientest.com\/wp-content\/uploads\/2020\/09\/Capture-de\u0301cran-2020-09-18-a\u0300-09.27.50-300x119.png 300w, https:\/\/datascientest.com\/wp-content\/uploads\/2020\/09\/Capture-de\u0301cran-2020-09-18-a\u0300-09.27.50-768x305.png 768w\" sizes=\"(max-width: 842px) 100vw, 842px\" \/>\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<div class=\"elementor-element elementor-element-d7754d2 elementor-widget elementor-widget-image\" data-id=\"d7754d2\" 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=\"2560\" height=\"675\" src=\"https:\/\/datascientest.com\/wp-content\/uploads\/2020\/09\/Sche\u0301ma-2.png\" class=\"attachment-full size-full wp-image-9811\" alt=\"\" srcset=\"https:\/\/datascientest.com\/wp-content\/uploads\/2020\/09\/Sche\u0301ma-2.png 2560w, https:\/\/datascientest.com\/wp-content\/uploads\/2020\/09\/Sche\u0301ma-2-300x79.png 300w, https:\/\/datascientest.com\/wp-content\/uploads\/2020\/09\/Sche\u0301ma-2-1024x270.png 1024w, https:\/\/datascientest.com\/wp-content\/uploads\/2020\/09\/Sche\u0301ma-2-768x202.png 768w, https:\/\/datascientest.com\/wp-content\/uploads\/2020\/09\/Sche\u0301ma-2-1536x405.png 1536w, https:\/\/datascientest.com\/wp-content\/uploads\/2020\/09\/Sche\u0301ma-2-2048x540.png 2048w\" sizes=\"(max-width: 2560px) 100vw, 2560px\" \/>\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<div class=\"elementor-element elementor-element-99a25c2 elementor-widget elementor-widget-heading\" data-id=\"99a25c2\" data-element_type=\"widget\" data-widget_type=\"heading.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t<h4 class=\"elementor-heading-title elementor-size-medium\">Mod\u00e8le<\/h4>\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-28f807b elementor-widget elementor-widget-text-editor\" data-id=\"28f807b\" 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-weight: 400;\">Le mod\u00e8le CBOW a des similitudes avec le mod\u00e8le de classification que l&rsquo;on vient d&rsquo;impl\u00e9menter dans la partie pr\u00e9c\u00e9dente. Notre mod\u00e8le sera compos\u00e9 des couches suivantes :<\/span><\/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-4527c68 elementor-widget elementor-widget-image\" data-id=\"4527c68\" 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<figure class=\"wp-caption\">\n\t\t\t\t\t\t\t\t\t\t<img decoding=\"async\" width=\"2560\" height=\"2385\" src=\"https:\/\/datascientest.com\/wp-content\/uploads\/2020\/09\/Sche\u0301ma-3.png\" class=\"attachment-full size-full wp-image-9821\" alt=\"cbow\" srcset=\"https:\/\/datascientest.com\/wp-content\/uploads\/2020\/09\/Sche\u0301ma-3.png 2560w, https:\/\/datascientest.com\/wp-content\/uploads\/2020\/09\/Sche\u0301ma-3-300x279.png 300w, https:\/\/datascientest.com\/wp-content\/uploads\/2020\/09\/Sche\u0301ma-3-1024x954.png 1024w, https:\/\/datascientest.com\/wp-content\/uploads\/2020\/09\/Sche\u0301ma-3-768x715.png 768w\" sizes=\"(max-width: 2560px) 100vw, 2560px\" \/>\t\t\t\t\t\t\t\t\t\t\t<figcaption class=\"widget-image-caption wp-caption-text\">Architecture CBOW<\/figcaption>\n\t\t\t\t\t\t\t\t\t\t<\/figure>\n\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-2ae7963 elementor-widget elementor-widget-text-editor\" data-id=\"2ae7963\" 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<ol><li style=\"font-weight: 400;\"><span style=\"font-weight: 400;\">La couche\u00a0<\/span><span style=\"font-weight: 400;\">Embedding<\/span><span style=\"font-weight: 400;\">\u00a0va <\/span><b>transformer chaque mot du contexte en vecteur d&#8217;embedding<\/b><span style=\"font-weight: 400;\">. La matrice\u00a0W\u00a0de l&#8217;embedding sera apprise au fur et \u00e0 mesure que le mod\u00e8le s&rsquo;entra\u00eene. Les dimensions r\u00e9sultantes sont : (lot, context_size, embedding).<\/span><\/li><li style=\"font-weight: 400;\"><span style=\"font-weight: 400;\">Ensuite, la couche\u00a0<\/span><span style=\"font-weight: 400;\">GlobalAveragePooling1D<\/span><span style=\"font-weight: 400;\">\u00a0permet de <\/span><b>sommer les diff\u00e9rents embedding<\/b><span style=\"font-weight: 400;\"> pour avoir une dimension en sortie (batch_size, embedding).<\/span><\/li><li style=\"font-weight: 400;\"><span style=\"font-weight: 400;\">Enfin, La couche\u00a0<\/span><span style=\"font-weight: 400;\">Dense<\/span><span style=\"font-weight: 400;\">\u00a0de taille\u00a0\u00ab\u00a0voc_size\u00a0\u00bb\u00a0<\/span><b>permet de pr\u00e9dire le mot cible<\/b><span style=\"font-weight: 400;\">.<\/span><\/li><\/ol>\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-2e70d66 elementor-widget elementor-widget-shortcode\" data-id=\"2e70d66\" 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\/15boehler\/806f56f30a993bb7c933acb1f23d9014.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-381c204 elementor-widget elementor-widget-text-editor\" data-id=\"381c204\" 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-weight: 400;\">La fonction de perte cross-entropy est g\u00e9n\u00e9ralement utilis\u00e9e pour entra\u00eener le mod\u00e8le\u00a0:<\/span><\/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-ea8ec46 elementor-widget elementor-widget-shortcode\" data-id=\"ea8ec46\" 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\/15boehler\/2e3c21701abdfe7dfec520cfc52c3465.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\t\t<\/div>\n\t\t<div class=\"elementor-element elementor-element-0524604 e-flex e-con-boxed e-con\" data-id=\"0524604\" data-element_type=\"container\" data-settings=\"{&quot;content_width&quot;:&quot;boxed&quot;}\">\n\t\t\t\t\t<div class=\"e-con-inner\">\n\t\t\t\t<div class=\"elementor-element elementor-element-ef7125f elementor-widget elementor-widget-menu-anchor\" data-id=\"ef7125f\" data-element_type=\"widget\" data-widget_type=\"menu-anchor.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t\t\t<div id=\"metrique\" class=\"elementor-menu-anchor\"><\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-12617c0 elementor-widget elementor-widget-heading\" data-id=\"12617c0\" data-element_type=\"widget\" data-widget_type=\"heading.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t<h3 class=\"elementor-heading-title elementor-size-default\">M\u00e9trique dans cette espace<\/h3>\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-78e8016 elementor-widget elementor-widget-text-editor\" data-id=\"78e8016\" 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-weight: 400;\">Maintenant que le mod\u00e8le est entra\u00een\u00e9, il peut \u00eatre int\u00e9ressant de comparer la distance entre les mots.<\/span><\/p><p><b>La \u00ab\u00a0Cosine similarity\u00a0\u00bb<\/b><span style=\"font-weight: 400;\">\u00a0est g\u00e9n\u00e9ralement utilis\u00e9e comme m\u00e9trique pour <\/span><b>mesurer la distance lorsque la norme des vecteurs n&rsquo;a pas d&rsquo;importance.<\/b><span style=\"font-weight: 400;\"> Cette m\u00e9trique capture la similitude entre deux mots. <\/span><\/p><blockquote><p><span style=\"font-weight: 400;\">Plus la cosine similarity\u00a0\u00bb est proche de 1, plus <\/span><b>les deux mots sont li\u00e9s.<\/b><\/p><\/blockquote>\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-58c1a1c elementor-widget elementor-widget-image\" data-id=\"58c1a1c\" 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=\"197\" src=\"https:\/\/datascientest.com\/wp-content\/uploads\/2020\/09\/word2vec.png\" class=\"attachment-full size-full wp-image-9910\" alt=\"word2vec\" srcset=\"https:\/\/datascientest.com\/wp-content\/uploads\/2020\/09\/word2vec.png 512w, https:\/\/datascientest.com\/wp-content\/uploads\/2020\/09\/word2vec-300x115.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<div class=\"elementor-element elementor-element-eaa08d8 elementor-widget elementor-widget-text-editor\" data-id=\"eaa08d8\" 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-weight: 400;\">Avec cette m\u00e9trique et dans ce sous-espace vectoriel, les 5 mots <\/span><b>les proches de \u00ab\u00a0body\u00a0\u00bb <\/b><span style=\"font-weight: 400;\">sont :<\/span><\/p><pre><span style=\"font-weight: 400;\">intestines\u00a0 --\u00a0 0.30548161268234253<\/span><br \/><span style=\"font-weight: 400;\">bodies\u00a0 --\u00a0 0.2691531181335449<\/span><br \/><span style=\"font-weight: 400;\">arm\u00a0 --\u00a0 0.24878980219364166<\/span><br \/><span style=\"font-weight: 400;\">chest\u00a0 --\u00a0 0.2261650413274765<\/span><br \/><span style=\"font-weight: 400;\">leg\u00a0 --\u00a0 0.2193179428577423<\/span><\/pre><p><span style=\"font-weight: 400;\">Les nombres ci-dessus repr\u00e9sentent les distances cos-similarit\u00e9 entre le mot \u00ab\u00a0body\u00a0\u00bb et des mots les plus proches.<\/span><\/p><p>Pour le mot \u00ab &lt;b&gt;zombie &lt;\/b&gt;\u00bb :<\/p><pre><span style=\"font-weight: 400;\">slasher\u00a0 --\u00a0 0.3172745406627655<\/span><br \/><span style=\"font-weight: 400;\">cannibal\u00a0 --\u00a0 0.28496912121772766<\/span><br \/><span style=\"font-weight: 400;\">zombies\u00a0 --\u00a0 0.2767203450202942<\/span><br \/><span style=\"font-weight: 400;\">horror\u00a0 --\u00a0 0.2607246935367584<\/span><br \/><span style=\"font-weight: 400;\">zombi\u00a0 --\u00a0 0.25878411531448364<\/span><\/pre><p>Pour le mot \u00ab &lt;b&gt;amazing &lt;\/b&gt;\u00bb :<\/p><pre><span style=\"font-weight: 400;\">brilliant\u00a0 --\u00a0 0.3372475802898407<\/span><br \/><span style=\"font-weight: 400;\">extraordinary\u00a0 --\u00a0 0.319326251745224<\/span><br \/><span style=\"font-weight: 400;\">great\u00a0 --\u00a0 0.29579296708106995<\/span><br \/><span style=\"font-weight: 400;\">breathtaking\u00a0 --\u00a0 0.2907085716724396<\/span><br \/><span style=\"font-weight: 400;\">fantastic\u00a0 --\u00a0 0.2871546149253845<\/span><\/pre><p>Pour le mot \u00ab&lt;b&gt; god&lt;\/b&gt; \u00bb :<\/p><pre><span style=\"font-weight: 400;\">heavens\u00a0 --\u00a0 0.268303781747818<\/span><br \/><span style=\"font-weight: 400;\">jesus\u00a0 --\u00a0 0.26807624101638794<\/span><br \/><span style=\"font-weight: 400;\">goodness\u00a0 --\u00a0 0.2618488669395447<\/span><br \/><span style=\"font-weight: 400;\">gods\u00a0 --\u00a0 0.24795521795749664<\/span><br \/><span style=\"font-weight: 400;\">doom\u00a0 --\u00a0 0.22242328524589539<\/span><\/pre>\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-5444c13 elementor-widget elementor-widget-heading\" data-id=\"5444c13\" data-element_type=\"widget\" data-widget_type=\"heading.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t<h4 class=\"elementor-heading-title elementor-size-medium\">Propri\u00e9t\u00e9s arithm\u00e9tiques<\/h4>\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-ef89c98 elementor-widget elementor-widget-text-editor\" data-id=\"ef89c98\" 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-weight: 400;\">Nous pouvons maintenant nous poser la question si notre sous-espace vectoriel des mots a des propri\u00e9t\u00e9s arithm\u00e9tiques. Prenons le c\u00e9l\u00e8bre exemple suivant :<\/span><\/p><p style=\"text-align: center;\"><span style=\"font-weight: 400;\">\ud835\udc3e\ud835\udc56\ud835\udc5b\ud835\udc54<\/span><span style=\"font-weight: 400;\">\u00a0\u2212\u00a0<\/span><span style=\"font-weight: 400;\">\ud835\udc40\ud835\udc4e\ud835\udc5b<\/span><span style=\"font-weight: 400;\">\u00a0+\u00a0<\/span><span style=\"font-weight: 400;\">\ud835\udc4a\ud835\udc5c\ud835\udc5a\ud835\udc4e\ud835\udc5b<\/span><span style=\"font-weight: 400;\">\u00a0=\u00a0<\/span><span style=\"font-size: 16px; font-weight: 400;\">\ud835\udc44\ud835\udc62\ud835\udc52\ud835\udc52\ud835\udc5b<\/span><\/p><p><span style=\"font-weight: 400;\">Dans cet exemple la <\/span><b>propri\u00e9t\u00e9 arithm\u00e9tique est la royaut\u00e9<\/b><span style=\"font-weight: 400;\">. Nous souhaitons v\u00e9rifier si cette propri\u00e9t\u00e9 va se propager \u00e0 \u00ab\u00a0woman\u00a0\u00bb. C\u2019est-\u00e0-dire que nous allons rechercher les mots les plus proches du vecteur suivant :<\/span><\/p><p style=\"text-align: center;\"><span style=\"font-weight: 400;\">arithmetic_vector = word2vec[index_word1] &#8211; word2vec[index_word2] + word2vec [index_word3]<\/span><\/p><p><span style=\"font-weight: 400;\">Ici, le mot \u00ab\u00a0queen\u00a0\u00bb n\u2019est pas beaucoup repr\u00e9sent\u00e9 dans notre jeu de donn\u00e9es, ce qui explique une mauvaise repr\u00e9sentation. Pour cette raison, nous allons plut\u00f4t nous <\/span><b>int\u00e9resser \u00e0 la propri\u00e9t\u00e9 de nombre<\/b><span style=\"font-weight: 400;\">\u00a0:<\/span><\/p><p style=\"text-align: center;\"><i><span style=\"font-weight: 400;\">Men<\/span><\/i><span style=\"font-weight: 400;\">\u00a0\u2212\u00a0<\/span><span style=\"font-weight: 400;\">\ud835\udc40\ud835\udc4e\ud835\udc5b<\/span><span style=\"font-weight: 400;\">\u00a0+\u00a0<\/span><span style=\"font-weight: 400;\">\ud835\udc4a\ud835\udc5c\ud835\udc5a\ud835\udc4e\ud835\udc5b<\/span><span style=\"font-weight: 400;\">\u00a0=\u00a0<\/span><span style=\"font-weight: 400;\">Women<\/span><\/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-61b7589 elementor-widget elementor-widget-text-editor\" data-id=\"61b7589\" 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>En utilisant l\u2019op\u00e9ration d\u00e9finie pr\u00e9c\u00e9demment et une m\u00e9trique de cos similarit\u00e9, <strong>les 5 mots les plus proches sont :<\/strong><\/p><pre><span style=\"font-weight: 400;\">women\u00a0 --\u00a0 0.2889893054962158<\/span><br \/><span style=\"font-weight: 400;\">females\u00a0 --\u00a0 0.272260844707489<\/span><br \/><span style=\"font-weight: 400;\">strangers\u00a0 --\u00a0 0.24558939039707184<\/span><br \/><span style=\"font-weight: 400;\">teens\u00a0 --\u00a0 0.2443128377199173<\/span><br \/><span style=\"font-weight: 400;\">daughters\u00a0 --\u00a0 0.24117740988731384<\/span><\/pre><p>Le r\u00e9sultat pour Zombies \u2212 Zombie + \ud835\udc4a\ud835\udc5c\ud835\udc5a\ud835\udc4e\ud835\udc5b :<\/p><pre><span style=\"font-weight: 400;\">women\u00a0 --\u00a0 0.2547883093357086<\/span><br \/><span style=\"font-weight: 400;\">females\u00a0 --\u00a0 0.23258551955223083<\/span><br \/><span style=\"font-weight: 400;\">ladies\u00a0 --\u00a0 0.22764989733695984<\/span><br \/><span style=\"font-weight: 400;\">stripper\u00a0 --\u00a0 0.22274985909461975<\/span><br \/><span style=\"font-weight: 400;\">develops\u00a0 --\u00a0 0.2202150821685791<\/span><br \/><br \/>Le r\u00e9sultat pour Men \u2212 Man + Soldier :<\/pre><pre><span style=\"font-weight: 400;\">soldiers\u00a0 --\u00a0 0.3547001779079437<\/span><br \/><span style=\"font-weight: 400;\">daughters\u00a0 --\u00a0 0.21896378695964813<\/span><br \/><span style=\"font-weight: 400;\">letters\u00a0 --\u00a0 0.21452251076698303<\/span><br \/><span style=\"font-weight: 400;\">backyard\u00a0 --\u00a0 0.21437858045101166<\/span><br \/><span style=\"font-weight: 400;\">veterans\u00a0 --\u00a0 0.21067838370800018<\/span><\/pre><p>Le r\u00e9sultat pour Zombies \u2212 Zombie + monster :<\/p><pre><span style=\"font-weight: 400;\">werewolves\u00a0 --\u00a0 0.2724993824958801<\/span><br \/><span style=\"font-weight: 400;\">monsters\u00a0 --\u00a0 0.25695472955703735<\/span><br \/><span style=\"font-weight: 400;\">creature\u00a0 --\u00a0 0.24453674256801605<\/span><br \/><span style=\"font-weight: 400;\">dragons\u00a0 --\u00a0 0.22363890707492828<\/span><br \/><span style=\"font-weight: 400;\">bloke\u00a0 --\u00a0 0.21858260035514832<\/span><\/pre>\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-38e4afb elementor-widget elementor-widget-text-editor\" data-id=\"38e4afb\" 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-weight: 400;\">Ici, le mot \u00ab\u00a0monsters\u00a0\u00bb et \u00ab\u00a0werewolves\u00a0\u00bb semblent \u00eatre tr\u00e8s proche dans le contexte de notre jeu de donn\u00e9es.<\/span><\/p><p><span style=\"font-weight: 400;\">En utilisant une ACP, nous pouvons trouver la dimension de la <\/span><b>propri\u00e9t\u00e9 de nombre\u00a0:<\/b><\/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-d8ca5e7 elementor-widget elementor-widget-image\" data-id=\"d8ca5e7\" 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=\"717\" height=\"683\" src=\"https:\/\/datascientest.com\/wp-content\/uploads\/2020\/09\/image_word2vec.png\" class=\"attachment-full size-full wp-image-9715\" alt=\"propri\u00e9t\u00e9 de nombre\" srcset=\"https:\/\/datascientest.com\/wp-content\/uploads\/2020\/09\/image_word2vec.png 717w, https:\/\/datascientest.com\/wp-content\/uploads\/2020\/09\/image_word2vec-300x286.png 300w\" sizes=\"(max-width: 717px) 100vw, 717px\" \/>\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<div class=\"elementor-element elementor-element-ff6ad55 elementor-widget elementor-widget-text-editor\" data-id=\"ff6ad55\" 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-weight: 400;\">En utilisant le m\u00eame raisonnement pour <\/span><b>capturer le genre\u00a0:<\/b><\/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-7033834 elementor-widget elementor-widget-image\" data-id=\"7033834\" 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=\"488\" src=\"https:\/\/datascientest.com\/wp-content\/uploads\/2020\/09\/word2vec_3.png\" class=\"attachment-full size-full wp-image-9908\" alt=\"word2vec\" srcset=\"https:\/\/datascientest.com\/wp-content\/uploads\/2020\/09\/word2vec_3.png 512w, https:\/\/datascientest.com\/wp-content\/uploads\/2020\/09\/word2vec_3-300x286.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<div class=\"elementor-element elementor-element-393547e elementor-widget elementor-widget-text-editor\" data-id=\"393547e\" 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-weight: 400;\">En utilisant le m\u00eame raisonnement entre les <\/span><b>verbes \u00e0 l\u2019infinitif et les verbes finissant par -ed\u00a0:<\/b><\/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-521f82c elementor-widget elementor-widget-image\" data-id=\"521f82c\" 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=\"489\" src=\"https:\/\/datascientest.com\/wp-content\/uploads\/2020\/09\/word2vec_2.png\" class=\"attachment-full size-full wp-image-9909\" alt=\"word2vec\" srcset=\"https:\/\/datascientest.com\/wp-content\/uploads\/2020\/09\/word2vec_2.png 512w, https:\/\/datascientest.com\/wp-content\/uploads\/2020\/09\/word2vec_2-300x287.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<div class=\"elementor-element elementor-element-deaf914 elementor-widget elementor-widget-text-editor\" data-id=\"deaf914\" 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-weight: 400;\">Le <\/span><b>\u00ab\u00a0word2vec embedding\u00a0\u00bb capture efficacement les propri\u00e9t\u00e9s\u00a0s\u00e9mantiques\u00a0et\u00a0arithm\u00e9tiques\u00a0d&rsquo;un mot<\/b><span style=\"font-weight: 400;\">. Il permet \u00e9galement <\/span><b>de r\u00e9duire la dimension du probl\u00e8me et par cons\u00e9quent la t\u00e2che d&rsquo;apprentissage<\/b><span style=\"font-weight: 400;\">.<\/span><\/p><p><span style=\"font-weight: 400;\">Nous pouvons nous imaginer utiliser l\u2019algorithme word2vec <\/span><b>pour pr\u00e9-entra\u00eener la matrice d\u2019embedding du mod\u00e8le de classification.<\/b><span style=\"font-weight: 400;\"> Par cons\u00e9quent, notre mod\u00e8le de classification aura une bien meilleure repr\u00e9sentation des mots lors de la phase d\u2019apprentissage des sentiments.<\/span><\/p><p><span style=\"font-weight: 400;\"><b>Envie de d\u00e9marrer une formation <\/b>plus d\u00e9taill\u00e9e sur les diff\u00e9rents aspects abord\u00e9s de cet article ?\u00a0<\/span><\/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-b8d9d1b elementor-align-center button-principale elementor-widget elementor-widget-button\" data-id=\"b8d9d1b\" data-element_type=\"widget\" data-widget_type=\"button.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t\t\t<div class=\"elementor-button-wrapper\">\n\t\t\t<a class=\"elementor-button elementor-button-link elementor-size-sm elementor-animation-grow\" href=\"https:\/\/datascientest.com\/formations-data\">\n\t\t\t\t\t\t<span class=\"elementor-button-content-wrapper\">\n\t\t\t\t\t\t<span class=\"elementor-button-text\">Explorez tous nos modules !<\/span>\n\t\t<\/span>\n\t\t\t\t\t<\/a>\n\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t\t\t<\/div>\n\t\t","protected":false},"excerpt":{"rendered":"<p>Le word embedding d\u00e9signe un ensemble de m\u00e9thode d\u2019apprentissage visant \u00e0 repr\u00e9senter les mots d\u2019un texte par des vecteurs de nombres r\u00e9els.&nbsp; Aujourd\u2019hui, nous allons vous pr\u00e9senter le 3\u00e8me volet de notre dossier NLP.&nbsp; Vous avez rat\u00e9 les premiers \u00e9pisodes ? Pas de panique les voici :&nbsp; Introduction au NLP Word embedding &#8211; Word2vec D\u00e9finition [&hellip;]<\/p>\n","protected":false},"author":78,"featured_media":9927,"comment_status":"open","ping_status":"closed","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>Word2vec : NLP &amp; Word Embedding - DataScientest<\/title>\n<meta name=\"description\" content=\"3\u00e8me volet de notre dossier NLP consacr\u00e9 au word embedding sur le fonctionnement et d\u2019impl\u00e9menter l&#039;algorithme Word2vec sur Python.\" \/>\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\/nlp-word-embedding-word2vec\" \/>\n<meta property=\"og:locale\" content=\"fr_FR\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"Word2vec : NLP &amp; Word Embedding\" \/>\n<meta property=\"og:url\" content=\"https:\/\/datascientest.com\/nlp-word-embedding-word2vec\" \/>\n<meta property=\"og:site_name\" content=\"Formation Data Science | DataScientest.com\" \/>\n<meta property=\"article:published_time\" content=\"2020-09-18T07:47:38+00:00\" \/>\n<meta property=\"article:modified_time\" content=\"2023-11-09T14:12:17+00:00\" \/>\n<meta property=\"og:image\" content=\"https:\/\/datascientest.com\/wp-content\/uploads\/2020\/09\/word2vec.jpg\" \/>\n\t<meta property=\"og:image:width\" content=\"1520\" \/>\n\t<meta property=\"og:image:height\" content=\"600\" \/>\n\t<meta property=\"og:image:type\" content=\"image\/jpeg\" \/>\n<meta name=\"author\" content=\"J\u00e9r\u00e9my Robert\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:label1\" content=\"\u00c9crit par\" \/>\n\t<meta name=\"twitter:data1\" content=\"J\u00e9r\u00e9my Robert\" \/>\n\t<meta name=\"twitter:label2\" content=\"Dur\u00e9e de lecture estim\u00e9e\" \/>\n\t<meta name=\"twitter:data2\" content=\"7 minutes\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\/\/schema.org\",\"@graph\":[{\"@type\":\"WebPage\",\"@id\":\"https:\/\/datascientest.com\/nlp-word-embedding-word2vec\",\"url\":\"https:\/\/datascientest.com\/nlp-word-embedding-word2vec\",\"name\":\"Word2vec : NLP &amp; Word Embedding - DataScientest\",\"isPartOf\":{\"@id\":\"https:\/\/datascientest.com\/#website\"},\"datePublished\":\"2020-09-18T07:47:38+00:00\",\"dateModified\":\"2023-11-09T14:12:17+00:00\",\"author\":{\"@id\":\"https:\/\/datascientest.com\/#\/schema\/person\/7af2361c48e24e79a8aa2360e6e74925\"},\"description\":\"3\u00e8me volet de notre dossier NLP consacr\u00e9 au word embedding sur le fonctionnement et d\u2019impl\u00e9menter l'algorithme Word2vec sur Python.\",\"inLanguage\":\"fr-FR\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\/\/datascientest.com\/nlp-word-embedding-word2vec\"]}]},{\"@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\/7af2361c48e24e79a8aa2360e6e74925\",\"name\":\"J\u00e9r\u00e9my Robert\",\"image\":{\"@type\":\"ImageObject\",\"inLanguage\":\"fr-FR\",\"@id\":\"https:\/\/datascientest.com\/#\/schema\/person\/image\/\",\"url\":\"https:\/\/datascientest.com\/wp-content\/uploads\/2023\/11\/jeremyrobert.jpeg\",\"contentUrl\":\"https:\/\/datascientest.com\/wp-content\/uploads\/2023\/11\/jeremyrobert.jpeg\",\"caption\":\"J\u00e9r\u00e9my Robert\"},\"description\":\"J\u00e9r\u00e9my Robert est actuellement Directeur p\u00e9dagogique B2B chez DataScientest.com. Avant de rejoindre DataScientest, il a travaill\u00e9 en tant que Consultant Business Intelligence chez DEAPDATA, o\u00f9 il a g\u00e9r\u00e9 des projets dans les domaines de l'assurance, bancaire et industrie. Il a \u00e9galement \u00e9t\u00e9 Co-fondateur et pr\u00e9sident de Tripzer. En tant que consultant ind\u00e9pendant, J\u00e9r\u00e9my a propos\u00e9 des formations sur MicroStrategy pour divers niveaux professionnels. Il est dipl\u00f4m\u00e9 de l'Universit\u00e9 Paris Cit\u00e9 avec un Master MIAGE, combinant comp\u00e9tences en informatique et gestion.\",\"sameAs\":[\"https:\/\/www.linkedin.com\/in\/jeremyrobert1\/\",\"Data Science\"],\"url\":\"https:\/\/datascientest.com\/author\/jeremyrobert\"}]}<\/script>\n<!-- \/ Yoast SEO Premium plugin. -->","yoast_head_json":{"title":"Word2vec : NLP &amp; Word Embedding - DataScientest","description":"3\u00e8me volet de notre dossier NLP consacr\u00e9 au word embedding sur le fonctionnement et d\u2019impl\u00e9menter l'algorithme Word2vec sur Python.","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\/nlp-word-embedding-word2vec","og_locale":"fr_FR","og_type":"article","og_title":"Word2vec : NLP &amp; Word Embedding","og_url":"https:\/\/datascientest.com\/nlp-word-embedding-word2vec","og_site_name":"Formation Data Science | DataScientest.com","article_published_time":"2020-09-18T07:47:38+00:00","article_modified_time":"2023-11-09T14:12:17+00:00","og_image":[{"width":1520,"height":600,"url":"https:\/\/datascientest.com\/wp-content\/uploads\/2020\/09\/word2vec.jpg","type":"image\/jpeg"}],"author":"J\u00e9r\u00e9my Robert","twitter_card":"summary_large_image","twitter_misc":{"\u00c9crit par":"J\u00e9r\u00e9my Robert","Dur\u00e9e de lecture estim\u00e9e":"7 minutes"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"WebPage","@id":"https:\/\/datascientest.com\/nlp-word-embedding-word2vec","url":"https:\/\/datascientest.com\/nlp-word-embedding-word2vec","name":"Word2vec : NLP &amp; Word Embedding - DataScientest","isPartOf":{"@id":"https:\/\/datascientest.com\/#website"},"datePublished":"2020-09-18T07:47:38+00:00","dateModified":"2023-11-09T14:12:17+00:00","author":{"@id":"https:\/\/datascientest.com\/#\/schema\/person\/7af2361c48e24e79a8aa2360e6e74925"},"description":"3\u00e8me volet de notre dossier NLP consacr\u00e9 au word embedding sur le fonctionnement et d\u2019impl\u00e9menter l'algorithme Word2vec sur Python.","inLanguage":"fr-FR","potentialAction":[{"@type":"ReadAction","target":["https:\/\/datascientest.com\/nlp-word-embedding-word2vec"]}]},{"@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\/7af2361c48e24e79a8aa2360e6e74925","name":"J\u00e9r\u00e9my Robert","image":{"@type":"ImageObject","inLanguage":"fr-FR","@id":"https:\/\/datascientest.com\/#\/schema\/person\/image\/","url":"https:\/\/datascientest.com\/wp-content\/uploads\/2023\/11\/jeremyrobert.jpeg","contentUrl":"https:\/\/datascientest.com\/wp-content\/uploads\/2023\/11\/jeremyrobert.jpeg","caption":"J\u00e9r\u00e9my Robert"},"description":"J\u00e9r\u00e9my Robert est actuellement Directeur p\u00e9dagogique B2B chez DataScientest.com. Avant de rejoindre DataScientest, il a travaill\u00e9 en tant que Consultant Business Intelligence chez DEAPDATA, o\u00f9 il a g\u00e9r\u00e9 des projets dans les domaines de l'assurance, bancaire et industrie. Il a \u00e9galement \u00e9t\u00e9 Co-fondateur et pr\u00e9sident de Tripzer. En tant que consultant ind\u00e9pendant, J\u00e9r\u00e9my a propos\u00e9 des formations sur MicroStrategy pour divers niveaux professionnels. Il est dipl\u00f4m\u00e9 de l'Universit\u00e9 Paris Cit\u00e9 avec un Master MIAGE, combinant comp\u00e9tences en informatique et gestion.","sameAs":["https:\/\/www.linkedin.com\/in\/jeremyrobert1\/","Data Science"],"url":"https:\/\/datascientest.com\/author\/jeremyrobert"}]}},"_links":{"self":[{"href":"https:\/\/datascientest.com\/wp-json\/wp\/v2\/posts\/9702"}],"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\/78"}],"replies":[{"embeddable":true,"href":"https:\/\/datascientest.com\/wp-json\/wp\/v2\/comments?post=9702"}],"version-history":[{"count":0,"href":"https:\/\/datascientest.com\/wp-json\/wp\/v2\/posts\/9702\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/datascientest.com\/wp-json\/wp\/v2\/media\/9927"}],"wp:attachment":[{"href":"https:\/\/datascientest.com\/wp-json\/wp\/v2\/media?parent=9702"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/datascientest.com\/wp-json\/wp\/v2\/categories?post=9702"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/datascientest.com\/wp-json\/wp\/v2\/tags?post=9702"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}