{"id":16103,"date":"2024-01-20T20:57:37","date_gmt":"2024-01-20T17:27:37","guid":{"rendered":"https:\/\/rasanegar.com\/blog\/python-%d8%a8%d8%b1%d8%a7%db%8c-nlp-%d8%a7%db%8c%d8%ac%d8%a7%d8%af-%d9%85%d8%af%d9%84%d9%87%d8%a7%db%8c-%d8%b7%d8%a8%d9%82%d9%87%d8%a8%d9%86%d8%af%db%8c-%da%86%d9%86%d8%af-%d8%af\/"},"modified":"2024-01-20T20:57:37","modified_gmt":"2024-01-20T17:27:37","slug":"python-%d8%a8%d8%b1%d8%a7%db%8c-nlp-%d8%a7%db%8c%d8%ac%d8%a7%d8%af-%d9%85%d8%af%d9%84%d9%87%d8%a7%db%8c-%d8%b7%d8%a8%d9%82%d9%87%d8%a8%d9%86%d8%af%db%8c-%da%86%d9%86%d8%af-%d8%af","status":"publish","type":"post","link":"https:\/\/rasanegaar.com\/blog\/python-%d8%a8%d8%b1%d8%a7%db%8c-nlp-%d8%a7%db%8c%d8%ac%d8%a7%d8%af-%d9%85%d8%af%d9%84%d9%87%d8%a7%db%8c-%d8%b7%d8%a8%d9%82%d9%87%d8%a8%d9%86%d8%af%db%8c-%da%86%d9%86%d8%af-%d8%af\/","title":{"rendered":"Python \u0628\u0631\u0627\u06cc NLP: \u0627\u06cc\u062c\u0627\u062f \u0645\u062f\u0644\u200c\u0647\u0627\u06cc \u0637\u0628\u0642\u0647\u200c\u0628\u0646\u062f\u06cc \u0686\u0646\u062f \u062f\u0627\u062f\u0647 \u0628\u0627 Keras"},"content":{"rendered":"<div id=\"ez-toc-container\" class=\"ez-toc-v2_0_85 counter-hierarchy ez-toc-counter ez-toc-custom ez-toc-container-direction\">\n<div class=\"ez-toc-title-container\"><p class=\"ez-toc-title\" style=\"cursor:inherit\">\u0633\u0631\u0641\u0635\u0644\u0647\u0627\u06cc \u0645\u0637\u0644\u0628<\/p>\n<\/div><nav><ul class='ez-toc-list ez-toc-list-level-1 ' ><li class='ez-toc-page-1 ez-toc-heading-level-2'><a class=\"ez-toc-link ez-toc-heading-1\" href=\"https:\/\/rasanegaar.com\/blog\/python-%d8%a8%d8%b1%d8%a7%db%8c-nlp-%d8%a7%db%8c%d8%ac%d8%a7%d8%af-%d9%85%d8%af%d9%84%d9%87%d8%a7%db%8c-%d8%b7%d8%a8%d9%82%d9%87%d8%a8%d9%86%d8%af%db%8c-%da%86%d9%86%d8%af-%d8%af\/#%d9%85%d8%ac%d9%85%d9%88%d8%b9%d9%87_%d8%af%d8%a7%d8%af%d9%87\" >\u0645\u062c\u0645\u0648\u0639\u0647 \u062f\u0627\u062f\u0647<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-2'><a class=\"ez-toc-link ez-toc-heading-2\" href=\"https:\/\/rasanegaar.com\/blog\/python-%d8%a8%d8%b1%d8%a7%db%8c-nlp-%d8%a7%db%8c%d8%ac%d8%a7%d8%af-%d9%85%d8%af%d9%84%d9%87%d8%a7%db%8c-%d8%b7%d8%a8%d9%82%d9%87%d8%a8%d9%86%d8%af%db%8c-%da%86%d9%86%d8%af-%d8%af\/#%d8%a7%db%8c%d8%ac%d8%a7%d8%af_%db%8c%da%a9_%d9%85%d8%af%d9%84_%d9%81%d9%82%d8%b7_%d8%a8%d8%a7_%d9%88%d8%b1%d9%88%d8%af%db%8c_%d9%87%d8%a7%db%8c_%d9%85%d8%aa%d9%86\" >\u0627\u06cc\u062c\u0627\u062f \u06cc\u06a9 \u0645\u062f\u0644 \u0641\u0642\u0637 \u0628\u0627 \u0648\u0631\u0648\u062f\u06cc \u0647\u0627\u06cc \u0645\u062a\u0646<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-2'><a class=\"ez-toc-link ez-toc-heading-3\" href=\"https:\/\/rasanegaar.com\/blog\/python-%d8%a8%d8%b1%d8%a7%db%8c-nlp-%d8%a7%db%8c%d8%ac%d8%a7%d8%af-%d9%85%d8%af%d9%84%d9%87%d8%a7%db%8c-%d8%b7%d8%a8%d9%82%d9%87%d8%a8%d9%86%d8%af%db%8c-%da%86%d9%86%d8%af-%d8%af\/#%d8%a7%db%8c%d8%ac%d8%a7%d8%af_%db%8c%da%a9_%d9%85%d8%af%d9%84_%d8%a8%d8%a7_%d9%88%d8%b1%d9%88%d8%af%db%8c_%d9%87%d8%a7%db%8c_%d9%85%d8%aa%d8%b9%d8%af%d8%af\" >\u0627\u06cc\u062c\u0627\u062f \u06cc\u06a9 \u0645\u062f\u0644 \u0628\u0627 \u0648\u0631\u0648\u062f\u06cc \u0647\u0627\u06cc \u0645\u062a\u0639\u062f\u062f<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-2'><a class=\"ez-toc-link ez-toc-heading-4\" href=\"https:\/\/rasanegaar.com\/blog\/python-%d8%a8%d8%b1%d8%a7%db%8c-nlp-%d8%a7%db%8c%d8%ac%d8%a7%d8%af-%d9%85%d8%af%d9%84%d9%87%d8%a7%db%8c-%d8%b7%d8%a8%d9%82%d9%87%d8%a8%d9%86%d8%af%db%8c-%da%86%d9%86%d8%af-%d8%af\/#%d8%a7%d9%81%da%a9%d8%a7%d8%b1_%d9%86%d9%87%d8%a7%db%8c%db%8c_%d9%88_%d8%a8%d9%87%d8%a8%d9%88%d8%af%d9%87%d8%a7\" >\u0627\u0641\u06a9\u0627\u0631 \u0646\u0647\u0627\u06cc\u06cc \u0648 \u0628\u0647\u0628\u0648\u062f\u0647\u0627<\/a><\/li><\/ul><\/nav><\/div>\n<span class=\"span-reading-time rt-reading-time\" style=\"display: block;\"><span class=\"rt-label rt-prefix\">\u0632\u0645\u0627\u0646 \u0644\u0627\u0632\u0645 \u0628\u0631\u0627\u06cc \u0645\u0637\u0627\u0644\u0639\u0647: <\/span> <span class=\"rt-time\"> 17<\/span> <span class=\"rt-label rt-postfix\">\u062f\u0642\u06cc\u0642\u0647<\/span><\/span><p> <br \/>\n<\/p>\n<div><noscript><\/noscript><\/p>\n<p>\u0627\u06cc\u0646 \u0647\u062c\u062f\u0647\u0645\u06cc\u0646 \u0645\u0642\u0627\u0644\u0647 \u0627\u0632 \u0633\u0631\u06cc \u0645\u0642\u0627\u0644\u0627\u062a \u0645\u0646 \u0627\u0633\u062a \u0631\u0648\u06cc \u067e\u0627\u06cc\u062a\u0648\u0646 \u0628\u0631\u0627\u06cc NLP.  \u062f\u0631 \u0645\u0642\u0627\u0644\u0647 \u0642\u0628\u0644\u06cc \u062e\u0648\u062f\u060c \u0631\u0648\u0634 \u0627\u06cc\u062c\u0627\u062f \u06cc\u06a9 \u0645\u062f\u0644 \u062a\u062d\u0644\u06cc\u0644 \u0627\u062d\u0633\u0627\u0633\u0627\u062a \u0641\u06cc\u0644\u0645 \u0645\u0628\u062a\u0646\u06cc \u0628\u0631 \u06cc\u0627\u062f\u06af\u06cc\u0631\u06cc \u0639\u0645\u06cc\u0642 \u0631\u0627 \u0628\u0627 \u0627\u0633\u062a\u0641\u0627\u062f\u0647 \u0627\u0632 \u067e\u0627\u06cc\u062a\u0648\u0646 \u062a\u0648\u0636\u06cc\u062d \u062f\u0627\u062f\u0645. <a rel=\"nofollow noopener\" target=\"_blank\" href=\"https:\/\/keras.io\/\">\u06a9\u0631\u0627\u0633<\/a> \u06a9\u062a\u0627\u0628\u062e\u0627\u0646\u0647  \u062f\u0631 \u0622\u0646 \u0645\u0642\u0627\u0644\u0647\u060c \u062f\u06cc\u062f\u06cc\u0645 \u06a9\u0647 \u0686\u06af\u0648\u0646\u0647 \u0645\u06cc\u200c\u062a\u0648\u0627\u0646\u06cc\u0645 \u062a\u062d\u0644\u06cc\u0644 \u0627\u062d\u0633\u0627\u0633\u0627\u062a \u0646\u0638\u0631\u0627\u062a \u06a9\u0627\u0631\u0628\u0631\u0627\u0646 \u0631\u0627 \u062f\u0631 \u0645\u0648\u0631\u062f \u0641\u06cc\u0644\u0645\u200c\u0647\u0627\u06cc \u0645\u062e\u062a\u0644\u0641 \u0627\u0646\u062c\u0627\u0645 \u062f\u0647\u06cc\u0645 \u0631\u0648\u06cc IMDB.  \u0645\u0627 \u0627\u0632 \u0645\u062a\u0646 \u0628\u0631\u0631\u0633\u06cc \u0628\u0631\u0627\u06cc \u067e\u06cc\u0634 \u0628\u06cc\u0646\u06cc \u0627\u062d\u0633\u0627\u0633\u0627\u062a \u0627\u0633\u062a\u0641\u0627\u062f\u0647 \u06a9\u0631\u062f\u06cc\u0645.<\/p>\n<p>\u0628\u0627 \u0627\u06cc\u0646 \u062d\u0627\u0644\u060c \u062f\u0631 \u06a9\u0627\u0631\u0647\u0627\u06cc \u0637\u0628\u0642\u0647 \u0628\u0646\u062f\u06cc \u0645\u062a\u0646\u060c \u0645\u06cc \u062a\u0648\u0627\u0646\u06cc\u0645 \u0627\u0632 \u0627\u0637\u0644\u0627\u0639\u0627\u062a \u063a\u06cc\u0631 \u0645\u062a\u0646\u06cc \u0646\u06cc\u0632 \u0628\u0631\u0627\u06cc \u0637\u0628\u0642\u0647 \u0628\u0646\u062f\u06cc \u0645\u062a\u0646 \u0627\u0633\u062a\u0641\u0627\u062f\u0647 \u06a9\u0646\u06cc\u0645.  \u0628\u0647 \u0639\u0646\u0648\u0627\u0646 \u0645\u062b\u0627\u0644\u060c \u062c\u0646\u0633\u06cc\u062a \u0645\u0645\u06a9\u0646 \u0627\u0633\u062a \u062a\u0623\u062b\u06cc\u0631 \u062f\u0627\u0634\u062a\u0647 \u0628\u0627\u0634\u062f \u0631\u0648\u06cc \u0627\u062d\u0633\u0627\u0633 \u0628\u0631\u0631\u0633\u06cc  \u0639\u0644\u0627\u0648\u0647 \u0628\u0631 \u0627\u06cc\u0646\u060c \u0645\u0644\u06cc\u062a \u0647\u0627 \u0645\u0645\u06a9\u0646 \u0627\u0633\u062a \u0628\u0631 \u0627\u0641\u06a9\u0627\u0631 \u0639\u0645\u0648\u0645\u06cc \u062f\u0631 \u0645\u0648\u0631\u062f \u06cc\u06a9 \u0641\u06cc\u0644\u0645 \u062e\u0627\u0635 \u062a\u0623\u062b\u06cc\u0631 \u0628\u06af\u0630\u0627\u0631\u0646\u062f.  \u0628\u0646\u0627\u0628\u0631\u0627\u06cc\u0646\u060c \u0627\u06cc\u0646 \u0627\u0637\u0644\u0627\u0639\u0627\u062a \u0645\u0631\u062a\u0628\u0637\u060c \u0647\u0645\u0686\u0646\u06cc\u0646 \u0628\u0647 \u0639\u0646\u0648\u0627\u0646 \u0641\u0631\u0627\u062f\u0627\u062f\u0647 \u0634\u0646\u0627\u062e\u062a\u0647 \u0645\u06cc \u0634\u0648\u062f\u060c \u0647\u0645\u0686\u0646\u06cc\u0646 \u0645\u06cc \u062a\u0648\u0627\u0646\u062f \u0628\u0631\u0627\u06cc \u0628\u0647\u0628\u0648\u062f \u062f\u0642\u062a \u0645\u062f\u0644 \u0622\u0645\u0627\u0631\u06cc \u0627\u0633\u062a\u0641\u0627\u062f\u0647 \u0634\u0648\u062f.<\/p>\n<p>\u062f\u0631 \u0627\u06cc\u0646 \u0645\u0642\u0627\u0644\u0647\u060c \u0645\u0627 \u0628\u0631 \u0627\u0633\u0627\u0633 \u0645\u0641\u0627\u0647\u06cc\u0645\u06cc \u06a9\u0647 \u062f\u0631 \u062f\u0648 \u0645\u0642\u0627\u0644\u0647 \u0627\u062e\u06cc\u0631 \u0645\u0637\u0627\u0644\u0639\u0647 \u06a9\u0631\u062f\u06cc\u0645\u060c \u0645\u06cc\u200c\u067e\u0631\u062f\u0627\u0632\u06cc\u0645 \u0648 \u062e\u0648\u0627\u0647\u06cc\u0645 \u062f\u06cc\u062f \u06a9\u0647 \u0686\u06af\u0648\u0646\u0647 \u0645\u06cc\u200c\u062a\u0648\u0627\u0646 \u06cc\u06a9 \u0633\u06cc\u0633\u062a\u0645 \u0637\u0628\u0642\u0647\u200c\u0628\u0646\u062f\u06cc \u0645\u062a\u0646 \u0627\u06cc\u062c\u0627\u062f \u06a9\u0631\u062f \u06a9\u0647 \u0646\u0638\u0631\u0627\u062a \u06a9\u0627\u0631\u0628\u0631\u0627\u0646 \u0631\u0627 \u062f\u0631 \u0645\u0648\u0631\u062f \u0645\u0634\u0627\u063a\u0644 \u0645\u062e\u062a\u0644\u0641 \u062f\u0631 \u06cc\u06a9\u06cc \u0627\u0632 \u0633\u0647 \u062f\u0633\u062a\u0647 \u0627\u0632 \u067e\u06cc\u0634 \u062a\u0639\u0631\u06cc\u0641\u200c\u0634\u062f\u0647 \u06cc\u0639\u0646\u06cc \u00ab\u062e\u0648\u0628\u00bb\u060c \u00ab\u0628\u062f\u00bb \u0637\u0628\u0642\u0647\u200c\u0628\u0646\u062f\u06cc \u06a9\u0646\u062f. &#8220;\u060c \u0648 &#8220;\u0645\u062a\u0648\u0633\u0637&#8221;.  \u0628\u0627 \u0627\u06cc\u0646 \u062d\u0627\u0644\u060c \u0639\u0644\u0627\u0648\u0647 \u0628\u0631 \u0645\u062a\u0646 \u0628\u0631\u0631\u0633\u06cc\u060c \u0627\u0632 \u0641\u0631\u0627\u062f\u0627\u062f\u0647 \u0645\u0631\u0628\u0648\u0637 \u0628\u0647 \u0628\u0631\u0631\u0633\u06cc \u0628\u0631\u0627\u06cc \u0627\u0646\u062c\u0627\u0645 \u0637\u0628\u0642\u0647\u200c\u0628\u0646\u062f\u06cc \u0627\u0633\u062a\u0641\u0627\u062f\u0647 \u062e\u0648\u0627\u0647\u06cc\u0645 \u06a9\u0631\u062f.  \u0627\u0632 \u0622\u0646\u062c\u0627\u06cc\u06cc \u06a9\u0647 \u0645\u0627 \u062f\u0648 \u0646\u0648\u0639 \u0648\u0631\u0648\u062f\u06cc \u0645\u062a\u0641\u0627\u0648\u062a \u062f\u0627\u0631\u06cc\u0645 \u06cc\u0639\u0646\u06cc \u0648\u0631\u0648\u062f\u06cc \u0645\u062a\u0646\u06cc \u0648 \u0648\u0631\u0648\u062f\u06cc \u0639\u062f\u062f\u06cc\u060c \u0628\u0627\u06cc\u062f \u06cc\u06a9 \u0645\u062f\u0644 \u0648\u0631\u0648\u062f\u06cc \u0686\u0646\u062f\u06af\u0627\u0646\u0647 \u0627\u06cc\u062c\u0627\u062f \u06a9\u0646\u06cc\u0645.  \u0645\u0627 \u0627\u0632 Keras Functional API \u0627\u0633\u062a\u0641\u0627\u062f\u0647 \u062e\u0648\u0627\u0647\u06cc\u0645 \u06a9\u0631\u062f \u0632\u06cc\u0631\u0627 \u0627\u0632 \u0686\u0646\u062f\u06cc\u0646 \u0648\u0631\u0648\u062f\u06cc \u0648 \u0686\u0646\u062f \u0645\u062f\u0644 \u062e\u0631\u0648\u062c\u06cc \u067e\u0634\u062a\u06cc\u0628\u0627\u0646\u06cc \u0645\u06cc \u06a9\u0646\u062f.<\/p>\n<p>\u067e\u0633 \u0627\u0632 \u0645\u0637\u0627\u0644\u0639\u0647 \u0627\u06cc\u0646 \u0645\u0642\u0627\u0644\u0647\u060c \u0645\u06cc\u200c\u062a\u0648\u0627\u0646\u06cc\u062f \u06cc\u06a9 \u0645\u062f\u0644 \u06cc\u0627\u062f\u06af\u06cc\u0631\u06cc \u0639\u0645\u06cc\u0642 \u062f\u0631 Keras \u0627\u06cc\u062c\u0627\u062f \u06a9\u0646\u06cc\u062f \u06a9\u0647 \u0642\u0627\u062f\u0631 \u0628\u0647 \u067e\u0630\u06cc\u0631\u0634 \u0686\u0646\u062f\u06cc\u0646 \u0648\u0631\u0648\u062f\u06cc\u060c \u0628\u0647 \u0647\u0645 \u067e\u06cc\u0648\u0633\u062a\u0646 \u062f\u0648 \u062e\u0631\u0648\u062c\u06cc \u0648 \u0633\u067e\u0633 \u0627\u0646\u062c\u0627\u0645 \u0637\u0628\u0642\u0647\u200c\u0628\u0646\u062f\u06cc \u06cc\u0627 \u0631\u06af\u0631\u0633\u06cc\u0648\u0646 \u0628\u0627 \u0627\u0633\u062a\u0641\u0627\u062f\u0647 \u0627\u0632 \u0648\u0631\u0648\u062f\u06cc \u062a\u062c\u0645\u06cc\u0639 \u0627\u0633\u062a.<\/p>\n<p>\u0642\u0628\u0644 \u0627\u0632 \u0627\u06cc\u0646\u06a9\u0647 \u0628\u0647 \u062c\u0632\u0626\u06cc\u0627\u062a \u0627\u06cc\u062c\u0627\u062f \u0686\u0646\u06cc\u0646 \u0645\u062f\u0644\u06cc \u0628\u067e\u0631\u062f\u0627\u0632\u06cc\u0645\u060c \u0627\u062c\u0627\u0632\u0647 \u062f\u0647\u06cc\u062f \u0627\u0628\u062a\u062f\u0627 \u0628\u0647 \u0637\u0648\u0631 \u062e\u0644\u0627\u0635\u0647 \u0645\u062c\u0645\u0648\u0639\u0647 \u062f\u0627\u062f\u0647 \u0627\u06cc \u0631\u0627 \u06a9\u0647 \u0642\u0631\u0627\u0631 \u0627\u0633\u062a \u0627\u0633\u062a\u0641\u0627\u062f\u0647 \u06a9\u0646\u06cc\u0645 \u0645\u0631\u0648\u0631 \u06a9\u0646\u06cc\u0645.<\/p>\n<h2 id=\"thedataset\"><span class=\"ez-toc-section\" id=\"%d9%85%d8%ac%d9%85%d9%88%d8%b9%d9%87_%d8%af%d8%a7%d8%af%d9%87\"><\/span>\u0645\u062c\u0645\u0648\u0639\u0647 \u062f\u0627\u062f\u0647<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<p>\u0645\u062c\u0645\u0648\u0639\u0647 \u062f\u0627\u062f\u0647 \u0647\u0627\u06cc \u0627\u06cc\u0646 \u0645\u0642\u0627\u0644\u0647 \u0631\u0627 \u0645\u06cc \u062a\u0648\u0627\u0646\u06cc\u062f \u0627\u0632 \u0627\u06cc\u0646\u062c\u0627 \u062f\u0627\u0646\u0644\u0648\u062f \u06a9\u0646\u06cc\u062f <a rel=\"nofollow noopener\" target=\"_blank\" href=\"https:\/\/www.kaggle.com\/yelp-dataset\/yelp-dataset\/version\/4#yelp_review.csv\" class=\"broken_link\">\u0644\u06cc\u0646\u06a9 \u06a9\u0627\u06af\u0644<\/a>.  \u0645\u062c\u0645\u0648\u0639\u0647 \u062f\u0627\u062f\u0647 \u0634\u0627\u0645\u0644 \u0686\u0646\u062f\u06cc\u0646 \u0641\u0627\u06cc\u0644 \u0627\u0633\u062a\u060c \u0627\u0645\u0627 \u0645\u0627 \u0641\u0642\u0637 \u0628\u0647 \u0622\u0646 \u0639\u0644\u0627\u0642\u0647 \u062f\u0627\u0631\u06cc\u0645 <code>yelp_review.csv<\/code> \u0641\u0627\u06cc\u0644.  \u0627\u06cc\u0646 \u0641\u0627\u06cc\u0644 \u062d\u0627\u0648\u06cc \u0628\u06cc\u0634 \u0627\u0632 5.2 \u0645\u06cc\u0644\u06cc\u0648\u0646 \u0628\u0631\u0631\u0633\u06cc \u062f\u0631 \u0645\u0648\u0631\u062f \u0645\u0634\u0627\u063a\u0644 \u0645\u062e\u062a\u0644\u0641 \u0627\u0632 \u062c\u0645\u0644\u0647 \u0631\u0633\u062a\u0648\u0631\u0627\u0646 \u0647\u0627\u060c \u06a9\u0627\u0641\u0647 \u0647\u0627\u060c \u062f\u0646\u062f\u0627\u0646\u067e\u0632\u0634\u06a9\u0627\u0646\u060c \u067e\u0632\u0634\u06a9\u0627\u0646\u060c \u0633\u0627\u0644\u0646 \u0647\u0627\u06cc \u0632\u06cc\u0628\u0627\u06cc\u06cc \u0648 \u063a\u06cc\u0631\u0647 \u0627\u0633\u062a. \u0628\u0631\u0627\u06cc \u0627\u0647\u062f\u0627\u0641 \u062e\u0648\u062f\u060c \u0645\u0627 \u0641\u0642\u0637 \u0627\u0632 50000 \u0631\u06a9\u0648\u0631\u062f \u0627\u0648\u0644 \u0628\u0631\u0627\u06cc \u0622\u0645\u0648\u0632\u0634 \u0645\u062f\u0644 \u062e\u0648\u062f \u0627\u0633\u062a\u0641\u0627\u062f\u0647 \u062e\u0648\u0627\u0647\u06cc\u0645 \u06a9\u0631\u062f.  \u0645\u062c\u0645\u0648\u0639\u0647 \u062f\u0627\u062f\u0647 \u0631\u0627 \u062f\u0631 \u062f\u0633\u062a\u06af\u0627\u0647 \u0645\u062d\u0644\u06cc \u062e\u0648\u062f \u062f\u0627\u0646\u0644\u0648\u062f \u06a9\u0646\u06cc\u062f.<\/p>\n<p>\u0627\u0648\u0644 \u0628\u06cc\u0627\u06cc\u06cc\u062f import \u062a\u0645\u0627\u0645 \u06a9\u062a\u0627\u0628\u062e\u0627\u0646\u0647 \u0647\u0627\u06cc\u06cc \u06a9\u0647 \u062f\u0631 \u0627\u06cc\u0646 \u0645\u0642\u0627\u0644\u0647 \u0642\u0628\u0644 \u0627\u0632 \u0648\u0627\u0631\u062f \u06a9\u0631\u062f\u0646 \u0645\u062c\u0645\u0648\u0639\u0647 \u062f\u0627\u062f\u0647 \u0627\u0633\u062a\u0641\u0627\u062f\u0647 \u062e\u0648\u0627\u0647\u06cc\u0645 \u06a9\u0631\u062f.<\/p>\n<pre><code class=\"hljs\"><span class=\"hljs-keyword\">from<\/span> numpy <span class=\"hljs-keyword\">import<\/span> array\n<span class=\"hljs-keyword\">from<\/span> keras.preprocessing.text <span class=\"hljs-keyword\">import<\/span> one_hot\n<span class=\"hljs-keyword\">from<\/span> keras.preprocessing.sequence <span class=\"hljs-keyword\">import<\/span> pad_sequences\n<span class=\"hljs-keyword\">from<\/span> keras.models <span class=\"hljs-keyword\">import<\/span> Sequential\n<span class=\"hljs-keyword\">from<\/span> keras.layers.core <span class=\"hljs-keyword\">import<\/span> Activation, Dropout, Dense\n<span class=\"hljs-keyword\">from<\/span> keras.layers <span class=\"hljs-keyword\">import<\/span> Flatten, LSTM\n<span class=\"hljs-keyword\">from<\/span> keras.layers <span class=\"hljs-keyword\">import<\/span> GlobalMaxPooling1D\n<span class=\"hljs-keyword\">from<\/span> keras.models <span class=\"hljs-keyword\">import<\/span> Model\n<span class=\"hljs-keyword\">from<\/span> keras.layers.embeddings <span class=\"hljs-keyword\">import<\/span> Embedding\n<span class=\"hljs-keyword\">from<\/span> sklearn.model_selection <span class=\"hljs-keyword\">import<\/span> train_test_split\n<span class=\"hljs-keyword\">from<\/span> keras.preprocessing.text <span class=\"hljs-keyword\">import<\/span> Tokenizer\n<span class=\"hljs-keyword\">from<\/span> keras.layers <span class=\"hljs-keyword\">import<\/span> Input\n<span class=\"hljs-keyword\">from<\/span> keras.layers.merge <span class=\"hljs-keyword\">import<\/span> Concatenate\n\n<span class=\"hljs-keyword\">import<\/span> pandas <span class=\"hljs-keyword\">as<\/span> pd\n<span class=\"hljs-keyword\">import<\/span> numpy <span class=\"hljs-keyword\">as<\/span> np\n<span class=\"hljs-keyword\">import<\/span> re\n<\/code><\/pre>\n<p>\u0628\u0647 \u0639\u0646\u0648\u0627\u0646 \u0627\u0648\u0644\u06cc\u0646 \u0642\u062f\u0645\u060c \u0628\u0627\u06cc\u062f \u0645\u062c\u0645\u0648\u0639\u0647 \u062f\u0627\u062f\u0647 \u0631\u0627 \u0628\u0627\u0631\u06af\u0630\u0627\u0631\u06cc \u06a9\u0646\u06cc\u0645.  \u0627\u0633\u06a9\u0631\u06cc\u067e\u062a \u0632\u06cc\u0631 \u0627\u06cc\u0646 \u06a9\u0627\u0631 \u0631\u0627 \u0627\u0646\u062c\u0627\u0645 \u0645\u06cc \u062f\u0647\u062f:<\/p>\n<pre><code class=\"hljs\">yelp_reviews = pd.read_csv(<span class=\"hljs-string\">\"\/content\/drive\/My Drive\/yelp_review_short.csv\"<\/span>)\n<\/code><\/pre>\n<p>\u0645\u062c\u0645\u0648\u0639\u0647 \u062f\u0627\u062f\u0647 \u0634\u0627\u0645\u0644 \u06cc\u06a9 \u0633\u062a\u0648\u0646 \u0627\u0633\u062a <code>Stars<\/code> \u06a9\u0647 \u062f\u0627\u0631\u0627\u06cc \u0631\u062a\u0628\u0647 \u0628\u0646\u062f\u06cc \u0628\u0631\u0627\u06cc \u0645\u0634\u0627\u063a\u0644 \u0645\u062e\u062a\u0644\u0641 \u0627\u0633\u062a.  \u0633\u062a\u0648\u0646 &#8220;\u0633\u062a\u0627\u0631\u0647 \u0647\u0627&#8221; \u0645\u06cc \u062a\u0648\u0627\u0646\u062f \u0645\u0642\u0627\u062f\u06cc\u0631\u06cc \u0628\u06cc\u0646 1 \u062a\u0627 5 \u062f\u0627\u0634\u062a\u0647 \u0628\u0627\u0634\u062f.  \u0645\u0627 \u06cc\u06a9 \u0633\u062a\u0648\u0646 \u062c\u062f\u06cc\u062f \u0627\u0636\u0627\u0641\u0647 \u0645\u06cc \u06a9\u0646\u06cc\u0645 <code>reviews_score<\/code> \u0628\u0647 \u0645\u062c\u0645\u0648\u0639\u0647 \u062f\u0627\u062f\u0647 \u0645\u0627  \u0627\u06af\u0631 \u0646\u0638\u0631 \u06a9\u0627\u0631\u0628\u0631 \u062f\u0627\u0631\u0627\u06cc \u0645\u0642\u062f\u0627\u0631 1 \u062f\u0631 \u0627\u0633\u062a <code>Stars<\/code> \u0633\u062a\u0648\u0646\u060c <code>reviews_score<\/code> \u0633\u062a\u0648\u0646 \u06cc\u06a9 \u0645\u0642\u062f\u0627\u0631 \u0631\u0634\u062a\u0647 \u0627\u06cc \u062e\u0648\u0627\u0647\u062f \u062f\u0627\u0634\u062a <code>bad<\/code>.  \u0627\u06af\u0631 \u0627\u0645\u062a\u06cc\u0627\u0632 2 \u06cc\u0627 3 \u0628\u0627\u0634\u062f <code>Stars<\/code> \u0633\u062a\u0648\u0646\u060c <code>reviews_score<\/code> \u0633\u062a\u0648\u0646 \u062d\u0627\u0648\u06cc \u06cc\u06a9 \u0645\u0642\u062f\u0627\u0631 \u062e\u0648\u0627\u0647\u062f \u0628\u0648\u062f <code>average<\/code>.  \u062f\u0631 \u0646\u0647\u0627\u06cc\u062a\u060c \u0631\u062a\u0628\u0647 \u0628\u0631\u0631\u0633\u06cc 4 \u06cc\u0627 5 \u062f\u0627\u0631\u0627\u06cc \u0645\u0642\u062f\u0627\u0631 \u0645\u0631\u0628\u0648\u0637\u0647 \u062e\u0648\u0627\u0647\u062f \u0628\u0648\u062f <code>good<\/code> \u062f\u0631 <code>reviews_score<\/code> \u0633\u062a\u0648\u0646<\/p>\n<p>\u0627\u0633\u06a9\u0631\u06cc\u067e\u062a \u0632\u06cc\u0631 \u0627\u06cc\u0646 \u067e\u06cc\u0634 \u067e\u0631\u062f\u0627\u0632\u0634 \u0631\u0627 \u0627\u0646\u062c\u0627\u0645 \u0645\u06cc \u062f\u0647\u062f:<\/p>\n<pre><code class=\"hljs\">bins = (<span class=\"hljs-number\">0<\/span>,<span class=\"hljs-number\">1<\/span>,<span class=\"hljs-number\">3<\/span>,<span class=\"hljs-number\">5<\/span>)\nreview_names = (<span class=\"hljs-string\">'bad'<\/span>, <span class=\"hljs-string\">'average'<\/span>, <span class=\"hljs-string\">'good'<\/span>)\nyelp_reviews(<span class=\"hljs-string\">'reviews_score'<\/span>) = pd.cut(yelp_reviews(<span class=\"hljs-string\">'stars'<\/span>), bins, labels=review_names)\n<\/code><\/pre>\n<p>\u062f\u0631 \u0645\u0631\u062d\u0644\u0647 \u0628\u0639\u062f\u060c \u062a\u0645\u0627\u0645 \u0645\u0642\u0627\u062f\u06cc\u0631 NULL \u0631\u0627 \u0627\u0632 \u062f\u06cc\u062a\u0627\u0641\u0631\u06cc\u0645 \u062e\u0648\u062f \u062d\u0630\u0641 \u0645\u06cc \u06a9\u0646\u06cc\u0645 \u0648 \u0627\u0631\u0627\u062f\u0647 \u0645\u06cc \u06a9\u0646\u06cc\u0645 print \u0634\u06a9\u0644 \u0648 \u0647\u062f\u0631 \u0645\u062c\u0645\u0648\u0639\u0647 \u062f\u0627\u062f\u0647<\/p>\n<pre><code class=\"hljs\">yelp_reviews.isnull().values.<span class=\"hljs-built_in\">any<\/span>()\n\n<span class=\"hljs-built_in\">print<\/span>(yelp_reviews.shape)\n\nyelp_reviews.head()\n<\/code><\/pre>\n<p>\u062f\u0631 \u062e\u0631\u0648\u062c\u06cc \u0645\u0634\u0627\u0647\u062f\u0647 \u062e\u0648\u0627\u0647\u06cc\u062f \u06a9\u0631\u062f <code>(50000,10)<\/code>\u060c \u0628\u0647 \u0627\u06cc\u0646 \u0645\u0639\u0646\u06cc \u06a9\u0647 \u0645\u062c\u0645\u0648\u0639\u0647 \u062f\u0627\u062f\u0647 \u0645\u0627 \u0634\u0627\u0645\u0644 50000 \u0631\u06a9\u0648\u0631\u062f \u0628\u0627 10 \u0633\u062a\u0648\u0646 \u0627\u0633\u062a.  \u0647\u062f\u0631 \u0627\u0632 <code>yelp_reviews<\/code> \u062f\u06cc\u062a\u0627\u0641\u0631\u06cc\u0645 \u0628\u0647 \u0634\u06a9\u0644 \u0632\u06cc\u0631 \u0627\u0633\u062a:<\/p>\n<p><img decoding=\"async\" class=\"img-responsive\" src=\"https:\/\/rasanegar.com\/blog\/wp-content\/uploads\/2024\/01\/python-for-nlp-creating-text-classification-model-with-multiple-inputs-in-keras-1.png\" alt=\"\u0633\u0631\" title=\"\"><\/p>\n<p>\u0645\u06cc \u062a\u0648\u0627\u0646\u06cc\u062f 10 \u0633\u062a\u0648\u0646\u06cc \u0631\u0627 \u06a9\u0647 \u062f\u06cc\u062a\u0627\u0641\u0631\u06cc\u0645 \u0645\u0627 \u0634\u0627\u0645\u0644 \u0645\u06cc \u0634\u0648\u062f\u060c \u0627\u0632 \u062c\u0645\u0644\u0647 \u0633\u062a\u0648\u0646 \u0647\u0627\u06cc\u06cc \u06a9\u0647 \u0628\u0647 \u062a\u0627\u0632\u06af\u06cc \u0627\u0636\u0627\u0641\u0647 \u0634\u062f\u0647 \u0627\u0633\u062a\u060c \u0645\u0634\u0627\u0647\u062f\u0647 \u06a9\u0646\u06cc\u062f <code>reviews_score<\/code> \u0633\u062a\u0648\u0646  \u0627\u06cc\u0646 <code>text<\/code> \u0633\u062a\u0648\u0646 \u062d\u0627\u0648\u06cc \u0645\u062a\u0646 \u0628\u0631\u0631\u0633\u06cc \u0627\u0633\u062a \u062f\u0631 \u062d\u0627\u0644\u06cc \u06a9\u0647 <code>useful<\/code> \u0633\u062a\u0648\u0646 \u062d\u0627\u0648\u06cc \u0645\u0642\u062f\u0627\u0631 \u0639\u062f\u062f\u06cc \u0627\u0633\u062a \u06a9\u0647 \u0646\u0634\u0627\u0646 \u062f\u0647\u0646\u062f\u0647 \u062a\u0639\u062f\u0627\u062f \u0627\u0641\u0631\u0627\u062f\u06cc \u0627\u0633\u062a \u06a9\u0647 \u0627\u06cc\u0646 \u0628\u0631\u0631\u0633\u06cc \u0631\u0627 \u0645\u0641\u06cc\u062f \u0645\u06cc \u062f\u0627\u0646\u0646\u062f.  \u0628\u0647 \u0637\u0648\u0631 \u0645\u0634\u0627\u0628\u0647\u060c <code>funny<\/code> \u0648 <code>cool<\/code> \u0633\u062a\u0648\u0646 \u0647\u0627 \u0634\u0627\u0645\u0644 \u062a\u0639\u062f\u0627\u062f \u0627\u0641\u0631\u0627\u062f\u06cc \u0627\u0633\u062a \u06a9\u0647 \u0646\u0638\u0631\u0627\u062a \u0631\u0627 \u067e\u06cc\u062f\u0627 \u06a9\u0631\u062f\u0647 \u0627\u0646\u062f <code>funny<\/code> \u06cc\u0627 <code>cool<\/code>\u060c \u0628\u0647 \u062a\u0631\u062a\u06cc\u0628.<\/p>\n<p>\u0628\u06cc\u0627\u06cc\u06cc\u062f \u0628\u0647 \u0637\u0648\u0631 \u062a\u0635\u0627\u062f\u0641\u06cc \u06cc\u06a9 \u0628\u0631\u0631\u0633\u06cc \u0631\u0627 \u0627\u0646\u062a\u062e\u0627\u0628 \u06a9\u0646\u06cc\u0645.  \u0627\u06af\u0631 \u0628\u0647 \u0628\u0631\u0631\u0633\u06cc \u0686\u0647\u0627\u0631\u0645 (\u0628\u0631\u0631\u0633\u06cc \u0628\u0627 \u0646\u0645\u0627\u06cc\u0647 3) \u0646\u06af\u0627\u0647 \u06a9\u0646\u06cc\u062f\u060c \u062f\u0627\u0631\u0627\u06cc 4 \u0633\u062a\u0627\u0631\u0647 \u0627\u0633\u062a \u0648 \u0627\u0632 \u0627\u06cc\u0646 \u0631\u0648 \u0628\u0647 \u0639\u0646\u0648\u0627\u0646 \u0639\u0644\u0627\u0645\u062a \u06af\u0630\u0627\u0631\u06cc \u0634\u062f\u0647 \u0627\u0633\u062a. <code>good<\/code>.  \u0628\u06cc\u0627\u06cc\u06cc\u062f \u0645\u062a\u0646 \u06a9\u0627\u0645\u0644 \u0627\u06cc\u0646 \u0628\u0631\u0631\u0633\u06cc \u0631\u0627 \u0645\u0634\u0627\u0647\u062f\u0647 \u06a9\u0646\u06cc\u0645:<\/p>\n<pre><code class=\"hljs\"><span class=\"hljs-built_in\">print<\/span>(yelp_reviews(<span class=\"hljs-string\">\"text\"<\/span>)(<span class=\"hljs-number\">3<\/span>))\n<\/code><\/pre>\n<p>\u062e\u0631\u0648\u062c\u06cc \u0628\u0647 \u0635\u0648\u0631\u062a \u0632\u06cc\u0631 \u0627\u0633\u062a:<\/p>\n<pre><code class=\"hljs\">Love coming here. Yes the place always needs the floor swept but when you give out  peanuts in the shell how won't it always be a bit dirty.\n\nThe food speaks for itself, so good. Burgers are made to order and the meat is put \u0631\u0648\u06cc the grill when you order your sandwich. Getting the small burger just means 1 patty, the regular is a 2 patty burger which is twice the deliciousness.\n\nGetting the Cajun fries adds a bit of spice to them and whatever size you order they always throw more fries (a lot more fries) into the bag.\n<\/code><\/pre>\n<p>\u0634\u0645\u0627 \u0628\u0647 \u0648\u0636\u0648\u062d \u0645\u06cc \u0628\u06cc\u0646\u06cc\u062f \u06a9\u0647 \u0627\u06cc\u0646 \u06cc\u06a9 \u0628\u0631\u0631\u0633\u06cc \u0645\u062b\u0628\u062a \u0627\u0633\u062a.<\/p>\n<p>\u062d\u0627\u0644\u0627 \u0628\u06cc\u0627\u06cc\u06cc\u062f \u062a\u0639\u062f\u0627\u062f \u0631\u0627 \u0631\u0633\u0645 \u06a9\u0646\u06cc\u0645 <code>good<\/code>\u060c <code>average<\/code>\u060c \u0648 <code>bad<\/code> \u0628\u0631\u0631\u0633\u06cc \u0647\u0627<\/p>\n<pre><code class=\"hljs\"><span class=\"hljs-keyword\">import<\/span> seaborn <span class=\"hljs-keyword\">as<\/span> sns\n\nsns.countplot(x=<span class=\"hljs-string\">'reviews_score'<\/span>, data=yelp_reviews)\n<\/code><\/pre>\n<p><img decoding=\"async\" class=\"img-responsive\" src=\"https:\/\/rasanegar.com\/blog\/wp-content\/uploads\/2024\/01\/python-for-nlp-creating-text-classification-model-with-multiple-inputs-in-keras-2.png\" alt=\"\u0633\u0631\" title=\"\"><\/p>\n<p>\u0627\u0632 \u0637\u0631\u062d \u0628\u0627\u0644\u0627 \u0645\u0634\u0647\u0648\u062f \u0627\u0633\u062a \u06a9\u0647 \u0627\u06a9\u062b\u0631 \u0628\u0631\u0631\u0633\u06cc \u0647\u0627 \u062e\u0648\u0628 \u0647\u0633\u062a\u0646\u062f \u0648 \u067e\u0633 \u0627\u0632 \u0622\u0646 \u0645\u06cc\u0627\u0646\u06af\u06cc\u0646 \u0646\u0642\u062f\u0647\u0627 \u0642\u0631\u0627\u0631 \u062f\u0627\u0631\u0646\u062f.  \u062a\u0639\u062f\u0627\u062f \u0646\u0642\u062f\u0647\u0627\u06cc \u0645\u0646\u0641\u06cc \u0628\u0633\u06cc\u0627\u0631 \u06a9\u0645 \u0627\u0633\u062a.<\/p>\n<p>\u0645\u0627 \u062f\u0627\u062f\u0647 \u0647\u0627\u06cc \u062e\u0648\u062f \u0631\u0627 \u0627\u0632 \u0642\u0628\u0644 \u067e\u0631\u062f\u0627\u0632\u0634 \u06a9\u0631\u062f\u0647 \u0627\u06cc\u0645 \u0648 \u0627\u06a9\u0646\u0648\u0646 \u062f\u0631 \u0627\u06cc\u0646 \u0645\u0642\u0627\u0644\u0647 \u0633\u0647 \u0645\u062f\u0644 \u0627\u06cc\u062c\u0627\u062f \u0645\u06cc \u06a9\u0646\u06cc\u0645.  \u0645\u062f\u0644 \u0627\u0648\u0644 \u0641\u0642\u0637 \u0627\u0632 \u0648\u0631\u0648\u062f\u06cc\u200c\u0647\u0627\u06cc \u0645\u062a\u0646\u06cc \u0628\u0631\u0627\u06cc \u067e\u06cc\u0634\u200c\u0628\u06cc\u0646\u06cc \u0627\u06cc\u0646\u06a9\u0647 \u0622\u06cc\u0627 \u0628\u0627\u0632\u0628\u06cc\u0646\u06cc \u0627\u0633\u062a \u0627\u0633\u062a\u0641\u0627\u062f\u0647 \u0645\u06cc\u200c\u06a9\u0646\u062f <code>good<\/code>\u060c <code>average<\/code>\u060c \u06cc\u0627 <code>bad<\/code>.  \u062f\u0631 \u0645\u062f\u0644 \u062f\u0648\u0645 \u0627\u0632 \u0645\u062a\u0646 \u0627\u0633\u062a\u0641\u0627\u062f\u0647 \u0646\u0645\u06cc \u06a9\u0646\u06cc\u0645.  \u0645\u0627 \u0641\u0642\u0637 \u0627\u0632 \u0627\u0637\u0644\u0627\u0639\u0627\u062a \u0645\u062a\u0627 \u0645\u0627\u0646\u0646\u062f <code>useful<\/code>\u060c <code>funny<\/code>\u060c \u0648 <code>cool<\/code> \u0628\u0631\u0627\u06cc \u067e\u06cc\u0634 \u0628\u06cc\u0646\u06cc \u0627\u062d\u0633\u0627\u0633 \u0628\u0631\u0631\u0633\u06cc  \u062f\u0631 \u0646\u0647\u0627\u06cc\u062a\u060c \u0645\u062f\u0644\u06cc \u0627\u06cc\u062c\u0627\u062f \u062e\u0648\u0627\u0647\u06cc\u0645 \u06a9\u0631\u062f \u06a9\u0647 \u0648\u0631\u0648\u062f\u06cc \u0647\u0627\u06cc \u0645\u062a\u0639\u062f\u062f\u06cc \u0645\u0627\u0646\u0646\u062f \u0645\u062a\u0646 \u0648 \u0627\u0637\u0644\u0627\u0639\u0627\u062a \u0645\u062a\u0627 \u0631\u0627 \u0628\u0631\u0627\u06cc \u0637\u0628\u0642\u0647 \u0628\u0646\u062f\u06cc \u0645\u062a\u0646 \u0645\u06cc \u067e\u0630\u06cc\u0631\u062f.<\/p>\n<h2 id=\"creatingamodelwithtextinputsonly\"><span class=\"ez-toc-section\" id=\"%d8%a7%db%8c%d8%ac%d8%a7%d8%af_%db%8c%da%a9_%d9%85%d8%af%d9%84_%d9%81%d9%82%d8%b7_%d8%a8%d8%a7_%d9%88%d8%b1%d9%88%d8%af%db%8c_%d9%87%d8%a7%db%8c_%d9%85%d8%aa%d9%86\"><\/span>\u0627\u06cc\u062c\u0627\u062f \u06cc\u06a9 \u0645\u062f\u0644 \u0641\u0642\u0637 \u0628\u0627 \u0648\u0631\u0648\u062f\u06cc \u0647\u0627\u06cc \u0645\u062a\u0646<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<p>\u0627\u0648\u0644\u06cc\u0646 \u06af\u0627\u0645\u060c \u062a\u0639\u0631\u06cc\u0641 \u062a\u0627\u0628\u0639\u06cc \u0627\u0633\u062a \u06a9\u0647 \u062f\u0627\u062f\u0647 \u0647\u0627\u06cc \u0645\u062a\u0646\u06cc \u0631\u0627 \u067e\u0627\u06a9 \u0645\u06cc \u06a9\u0646\u062f.<\/p>\n<pre><code class=\"hljs\"><span class=\"hljs-function\"><span class=\"hljs-keyword\">def<\/span> <span class=\"hljs-title\">preprocess_text<\/span>(<span class=\"hljs-params\">sen<\/span>):<\/span>\n\n    \n    sentence = re.sub(<span class=\"hljs-string\">'(^a-zA-Z)'<\/span>, <span class=\"hljs-string\">' '<\/span>, sen)\n\n    \n    sentence = re.sub(<span class=\"hljs-string\">r\"\\s+(a-zA-Z)\\s+\"<\/span>, <span class=\"hljs-string\">' '<\/span>, sentence)\n\n    \n    sentence = re.sub(<span class=\"hljs-string\">r'\\s+'<\/span>, <span class=\"hljs-string\">' '<\/span>, sentence)\n\n    <span class=\"hljs-keyword\">return<\/span> sentence\n<\/code><\/pre>\n<p>\u0627\u0632 \u0622\u0646\u062c\u0627\u06cc\u06cc \u06a9\u0647 \u062f\u0631 \u0627\u06cc\u0646 \u0645\u062f\u0644 \u0641\u0642\u0637 \u0627\u0632 \u0645\u062a\u0646 \u0627\u0633\u062a\u0641\u0627\u062f\u0647 \u0645\u06cc \u06a9\u0646\u06cc\u0645\u060c \u062a\u0645\u0627\u0645 \u0628\u0631\u0631\u0633\u06cc \u0647\u0627\u06cc \u0645\u062a\u0646 \u0631\u0627 \u0641\u06cc\u0644\u062a\u0631 \u06a9\u0631\u062f\u0647 \u0648 \u062f\u0631 \u0644\u06cc\u0633\u062a \u0630\u062e\u06cc\u0631\u0647 \u0645\u06cc \u06a9\u0646\u06cc\u0645.  \u0645\u0631\u0648\u0631\u0647\u0627\u06cc \u0645\u062a\u0646 \u0628\u0627 \u0627\u0633\u062a\u0641\u0627\u062f\u0647 \u0627\u0632 <code>preprocess_text<\/code> \u062a\u0627\u0628\u0639\u060c \u06a9\u0647 \u0639\u0644\u0627\u0626\u0645 \u0646\u06af\u0627\u0631\u0634\u06cc \u0648 \u0627\u0639\u062f\u0627\u062f \u0631\u0627 \u0627\u0632 \u0645\u062a\u0646 \u062d\u0630\u0641 \u0645\u06cc \u06a9\u0646\u062f.<\/p>\n<pre><code class=\"hljs\">X = ()\nsentences = <span class=\"hljs-built_in\">list<\/span>(yelp_reviews(<span class=\"hljs-string\">\"text\"<\/span>))\n<span class=\"hljs-keyword\">for<\/span> sen <span class=\"hljs-keyword\">in<\/span> sentences:\n    X.append(preprocess_text(sen))\n\ny = yelp_reviews(<span class=\"hljs-string\">'reviews_score'<\/span>)\n<\/code><\/pre>\n<p>\u0645\u0627 <code>X<\/code> \u0645\u062a\u063a\u06cc\u0631 \u062f\u0631 \u0627\u06cc\u0646\u062c\u0627 \u0634\u0627\u0645\u0644 \u0628\u0631\u0631\u0633\u06cc \u0645\u062a\u0646 \u0627\u0633\u062a \u062f\u0631 \u062d\u0627\u0644\u06cc \u06a9\u0647 <code>y<\/code> \u0645\u062a\u063a\u06cc\u0631 \u0634\u0627\u0645\u0644 \u0645\u0648\u0627\u0631\u062f \u0645\u0631\u0628\u0648\u0637\u0647 \u0627\u0633\u062a <code>reviews_score<\/code> \u0627\u0631\u0632\u0634 \u0647\u0627\u06cc.  \u0627\u06cc\u0646 <code>reviews_score<\/code> \u0633\u062a\u0648\u0646 \u062f\u0627\u0631\u0627\u06cc \u062f\u0627\u062f\u0647 \u062f\u0631 \u0642\u0627\u0644\u0628 \u0645\u062a\u0646 \u0627\u0633\u062a.  \u0645\u0627 \u0628\u0627\u06cc\u062f \u0645\u062a\u0646 \u0631\u0627 \u0628\u0647 \u06cc\u06a9 \u0628\u0631\u062f\u0627\u0631 \u0631\u0645\u0632\u06af\u0630\u0627\u0631\u06cc \u0634\u062f\u0647 \u06cc\u06a9 \u062f\u0627\u063a \u062a\u0628\u062f\u06cc\u0644 \u06a9\u0646\u06cc\u0645.  \u0645\u0627 \u0645\u06cc \u062a\u0648\u0627\u0646\u06cc\u0645 \u0627\u0633\u062a\u0641\u0627\u062f\u0647 \u06a9\u0646\u06cc\u0645 <code>to_categorical<\/code> \u0631\u0648\u0634 \u0627\u0632 <code>keras.utils<\/code> \u0645\u062f\u0648\u0644.  \u0628\u0627 \u0627\u06cc\u0646 \u062d\u0627\u0644\u060c \u0627\u0628\u062a\u062f\u0627 \u0628\u0627\u06cc\u062f \u0645\u062a\u0646 \u0631\u0627 \u0628\u0627 \u0627\u0633\u062a\u0641\u0627\u062f\u0647 \u0627\u0632 \u0628\u0631\u0686\u0633\u0628 \u0628\u0647 \u0639\u062f\u062f \u0635\u062d\u06cc\u062d \u062a\u0628\u062f\u06cc\u0644 \u06a9\u0646\u06cc\u0645 <code>LabelEncoder<\/code> \u062a\u0627\u0628\u0639 \u0627\u0632 <code>sklearn.preprocessing<\/code> \u0645\u062f\u0648\u0644.<\/p>\n<pre><code class=\"hljs\"><span class=\"hljs-keyword\">from<\/span> sklearn <span class=\"hljs-keyword\">import<\/span> preprocessing\n\n\nlabel_encoder = preprocessing.LabelEncoder()\n\n\ny = label_encoder.fit_transform(y)\n<\/code><\/pre>\n<p>\u0628\u06cc\u0627\u06cc\u06cc\u062f \u0627\u06a9\u0646\u0648\u0646 \u062f\u0627\u062f\u0647 \u0647\u0627\u06cc \u062e\u0648\u062f \u0631\u0627 \u0628\u0647 \u0645\u062c\u0645\u0648\u0639\u0647 \u0647\u0627\u06cc \u0622\u0632\u0645\u0627\u06cc\u0634\u06cc \u0648 \u0622\u0645\u0648\u0632\u0634\u06cc \u062a\u0642\u0633\u06cc\u0645 \u06a9\u0646\u06cc\u0645:<\/p>\n<pre><code class=\"hljs\">X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=<span class=\"hljs-number\">0.20<\/span>, random_state=<span class=\"hljs-number\">42<\/span>)\n<\/code><\/pre>\n<p>\u0627\u06a9\u0646\u0648\u0646 \u0645\u06cc\u200c\u062a\u0648\u0627\u0646\u06cc\u0645 \u0628\u0631\u0686\u0633\u0628\u200c\u0647\u0627\u06cc \u0622\u0645\u0648\u0632\u0634\u06cc \u0648 \u0622\u0632\u0645\u0627\u06cc\u0634\u06cc \u0631\u0627 \u0628\u0647 \u0628\u0631\u062f\u0627\u0631\u0647\u0627\u06cc \u0631\u0645\u0632\u06af\u0630\u0627\u0631\u06cc \u0634\u062f\u0647 \u06cc\u06a9\u200c\u0637\u0631\u0641\u0647 \u062a\u0628\u062f\u06cc\u0644 \u06a9\u0646\u06cc\u0645:<\/p>\n<pre><code class=\"hljs\"><span class=\"hljs-keyword\">from<\/span> keras.utils <span class=\"hljs-keyword\">import<\/span> to_categorical\ny_train = to_categorical(y_train)\ny_test = to_categorical(y_test)\n<\/code><\/pre>\n<p>\u062f\u0631 \u0645\u0642\u0627\u0644\u0647 \u0627\u0645 \u062a\u0648\u0636\u06cc\u062d \u062f\u0627\u062f\u0645 \u0631\u0648\u06cc \u062a\u0639\u0628\u06cc\u0647\u200c\u0647\u0627\u06cc \u06a9\u0644\u0645\u0647\u200c\u0627\u06cc \u06a9\u0647 \u062f\u0627\u062f\u0647\u200c\u0647\u0627\u06cc \u0645\u062a\u0646\u06cc \u0628\u0627\u06cc\u062f \u0628\u0647 \u0646\u0648\u0639\u06cc \u0634\u06a9\u0644 \u0639\u062f\u062f\u06cc \u062a\u0628\u062f\u06cc\u0644 \u0634\u0648\u0646\u062f \u062a\u0627 \u0628\u062a\u0648\u0627\u0646\u0646\u062f \u062a\u0648\u0633\u0637 \u0627\u0644\u06af\u0648\u0631\u06cc\u062a\u0645\u200c\u0647\u0627\u06cc \u0622\u0645\u0627\u0631\u06cc \u0645\u0627\u0646\u0646\u062f \u0645\u062f\u0644\u200c\u0647\u0627\u06cc \u06cc\u0627\u062f\u06af\u06cc\u0631\u06cc \u0645\u0627\u0634\u06cc\u0646\u06cc \u0648 \u0639\u0645\u06cc\u0642 \u0627\u0633\u062a\u0641\u0627\u062f\u0647 \u0634\u0648\u0646\u062f.  \u06cc\u06a9\u06cc \u0627\u0632 \u0631\u0627\u0647 \u0647\u0627\u06cc \u062a\u0628\u062f\u06cc\u0644 \u0645\u062a\u0646 \u0628\u0647 \u0627\u0639\u062f\u0627\u062f \u0627\u0632 \u0637\u0631\u06cc\u0642 \u062c\u0627\u0633\u0627\u0632\u06cc \u06a9\u0644\u0645\u0647 \u0627\u0633\u062a.  \u0627\u06af\u0631 \u0627\u0632 \u0631\u0648\u0634 \u067e\u06cc\u0627\u062f\u0647 \u0633\u0627\u0632\u06cc \u062c\u0627\u0633\u0627\u0632\u06cc \u06a9\u0644\u0645\u0647 \u0627\u0632 \u0637\u0631\u06cc\u0642 Keras \u0628\u06cc \u0627\u0637\u0644\u0627\u0639 \u0647\u0633\u062a\u06cc\u062f\u060c \u0628\u0647 \u0634\u062f\u062a \u062a\u0648\u0635\u06cc\u0647 \u0645\u06cc \u06a9\u0646\u0645 \u0642\u0628\u0644 \u0627\u0632 \u062d\u0631\u06a9\u062a \u0627\u06cc\u0646 \u0645\u0642\u0627\u0644\u0647 \u0631\u0627 \u0628\u062e\u0648\u0627\u0646\u06cc\u062f. \u0631\u0648\u06cc \u0628\u0647 \u0628\u062e\u0634 \u0647\u0627\u06cc \u0628\u0639\u062f\u06cc \u06a9\u062f.<\/p>\n<p>\u0627\u0648\u0644\u06cc\u0646 \u06af\u0627\u0645 \u062f\u0631 \u062c\u0627\u0633\u0627\u0632\u06cc \u06a9\u0644\u0645\u0627\u062a\u060c \u062a\u0628\u062f\u06cc\u0644 \u06a9\u0644\u0645\u0627\u062a \u0628\u0647 \u0646\u0645\u0627\u06cc\u0647 \u0647\u0627\u06cc \u0639\u062f\u062f\u06cc \u0645\u0631\u0628\u0648\u0637 \u0628\u0647 \u0622\u0646\u0647\u0627\u0633\u062a.  \u0628\u0631\u0627\u06cc \u0627\u06cc\u0646 \u06a9\u0627\u0631 \u0645\u06cc \u062a\u0648\u0627\u0646\u06cc\u0645 \u0627\u0632 <code>Tokenizer<\/code> \u06a9\u0644\u0627\u0633 \u0627\u0632 <code>Keras.preprocessing.text<\/code> \u0645\u062f\u0648\u0644.<\/p>\n<pre><code class=\"hljs\">tokenizer = Tokenizer(num_words=<span class=\"hljs-number\">5000<\/span>)\ntokenizer.fit_on_texts(X_train)\n\nX_train = tokenizer.texts_to_sequences(X_train)\nX_test = tokenizer.texts_to_sequences(X_test)\n<\/code><\/pre>\n<p>\u062c\u0645\u0644\u0627\u062a \u0645\u06cc \u062a\u0648\u0627\u0646\u0646\u062f \u0637\u0648\u0644 \u0647\u0627\u06cc \u0645\u062a\u0641\u0627\u0648\u062a\u06cc \u062f\u0627\u0634\u062a\u0647 \u0628\u0627\u0634\u0646\u062f\u060c \u0648 \u0628\u0646\u0627\u0628\u0631\u0627\u06cc\u0646 \u062f\u0646\u0628\u0627\u0644\u0647 \u0647\u0627\u06cc\u06cc \u06a9\u0647 \u062a\u0648\u0633\u0637 the \u0628\u0631\u06af\u0631\u062f\u0627\u0646\u062f\u0647 \u0645\u06cc \u0634\u0648\u0646\u062f <code>Tokenizer<\/code> \u06a9\u0644\u0627\u0633 \u0646\u06cc\u0632 \u0627\u0632 \u0637\u0648\u0644 \u0647\u0627\u06cc \u0645\u062a\u063a\u06cc\u0631 \u062a\u0634\u06a9\u06cc\u0644 \u0634\u062f\u0647 \u0627\u0633\u062a.  \u0645\u0627 \u0645\u0634\u062e\u0635 \u0645\u06cc \u06a9\u0646\u06cc\u0645 \u06a9\u0647 \u062d\u062f\u0627\u06a9\u062b\u0631 \u0637\u0648\u0644 \u062f\u0646\u0628\u0627\u0644\u0647 200 \u0628\u0627\u0634\u062f (\u0627\u0644\u0628\u062a\u0647 \u0645\u06cc \u062a\u0648\u0627\u0646\u06cc\u062f \u0647\u0631 \u0639\u062f\u062f\u06cc \u0631\u0627 \u0627\u0645\u062a\u062d\u0627\u0646 \u06a9\u0646\u06cc\u062f).  \u0628\u0631\u0627\u06cc \u062c\u0645\u0644\u0627\u062a\u06cc \u06a9\u0647 \u0637\u0648\u0644 \u0622\u0646\u0647\u0627 \u06a9\u0645\u062a\u0631 \u0627\u0632 200 \u0627\u0633\u062a\u060c \u0634\u0627\u062e\u0635 \u0647\u0627\u06cc \u0628\u0627\u0642\u06cc\u0645\u0627\u0646\u062f\u0647 \u0628\u0627 \u0635\u0641\u0631 \u067e\u0631 \u0645\u06cc \u0634\u0648\u0646\u062f.  \u0628\u0631\u0627\u06cc \u062c\u0645\u0644\u0627\u062a\u06cc \u06a9\u0647 \u0637\u0648\u0644 \u0622\u0646\u0647\u0627 \u0628\u06cc\u0634\u062a\u0631 \u0627\u0632 200 \u0628\u0627\u0634\u062f\u060c \u0634\u0627\u062e\u0635 \u0647\u0627\u06cc \u0628\u0627\u0642\u06cc\u0645\u0627\u0646\u062f\u0647 \u06a9\u0648\u062a\u0627\u0647 \u0645\u06cc \u0634\u0648\u0646\u062f.<\/p>\n<p>\u0628\u0647 \u0627\u0633\u06a9\u0631\u06cc\u067e\u062a \u0632\u06cc\u0631 \u0646\u06af\u0627\u0647 \u06a9\u0646\u06cc\u062f:<\/p>\n<pre><code class=\"hljs\">vocab_size = <span class=\"hljs-built_in\">len<\/span>(tokenizer.word_index) + <span class=\"hljs-number\">1<\/span>\n\nmaxlen = <span class=\"hljs-number\">200<\/span>\n\nX_train = pad_sequences(X_train, padding=<span class=\"hljs-string\">'post'<\/span>, maxlen=maxlen)\nX_test = pad_sequences(X_test, padding=<span class=\"hljs-string\">'post'<\/span>, maxlen=maxlen)\n<\/code><\/pre>\n<p>\u0628\u0639\u062f\u060c \u0645\u0627 \u0628\u0627\u06cc\u062f \u062f\u0627\u062e\u0644\u06cc \u0631\u0627 \u0628\u0627\u0631\u06af\u0630\u0627\u0631\u06cc \u06a9\u0646\u06cc\u0645 <a rel=\"nofollow noopener\" target=\"_blank\" href=\"https:\/\/nlp.stanford.edu\/projects\/glove\/\">\u062f\u0633\u062a\u06a9\u0634<\/a> \u062c\u0627\u0633\u0627\u0632\u06cc \u06a9\u0644\u0645\u0627\u062a<\/p>\n<pre><code class=\"hljs\"><span class=\"hljs-keyword\">from<\/span> numpy <span class=\"hljs-keyword\">import<\/span> array\n<span class=\"hljs-keyword\">from<\/span> numpy <span class=\"hljs-keyword\">import<\/span> asarray\n<span class=\"hljs-keyword\">from<\/span> numpy <span class=\"hljs-keyword\">import<\/span> zeros\n\nembeddings_dictionary = <span class=\"hljs-built_in\">dict<\/span>()\n\n<span class=\"hljs-keyword\">for<\/span> line <span class=\"hljs-keyword\">in<\/span> glove_file:\n    records = line.split()\n    word = records(<span class=\"hljs-number\">0<\/span>)\n    vector_dimensions = asarray(records(<span class=\"hljs-number\">1<\/span>:), dtype=<span class=\"hljs-string\">'float32'<\/span>)\n    embeddings_dictionary (word) = vector_dimensions\n\nglove_file.close()\n<\/code><\/pre>\n<p>\u062f\u0631 \u0646\u0647\u0627\u06cc\u062a \u06cc\u06a9 \u0645\u0627\u062a\u0631\u06cc\u0633 \u062c\u0627\u0633\u0627\u0632\u06cc \u0627\u06cc\u062c\u0627\u062f \u0645\u06cc \u06a9\u0646\u06cc\u0645 \u06a9\u0647 \u062f\u0631 \u0622\u0646 \u0631\u062f\u06cc\u0641 \u0647\u0627 \u0628\u0631\u0627\u0628\u0631 \u0628\u0627 \u062a\u0639\u062f\u0627\u062f \u06a9\u0644\u0645\u0627\u062a \u0645\u0648\u062c\u0648\u062f \u062f\u0631 \u0648\u0627\u0698\u06af\u0627\u0646 (\u0628\u0647 \u0627\u0636\u0627\u0641\u0647 1) \u062e\u0648\u0627\u0647\u062f \u0628\u0648\u062f.  \u062a\u0639\u062f\u0627\u062f \u0633\u062a\u0648\u0646 \u0647\u0627 100 \u062e\u0648\u0627\u0647\u062f \u0628\u0648\u062f \u0632\u06cc\u0631\u0627 \u0647\u0631 \u06a9\u0644\u0645\u0647 \u062f\u0631 \u062c\u0627\u0633\u0627\u0632\u06cc \u06a9\u0644\u0645\u0647 GloVe \u06a9\u0647 \u0628\u0627\u0631\u06af\u0630\u0627\u0631\u06cc \u06a9\u0631\u062f\u06cc\u0645 \u0628\u0647 \u0639\u0646\u0648\u0627\u0646 \u06cc\u06a9 \u0628\u0631\u062f\u0627\u0631 100 \u0628\u0639\u062f\u06cc \u0646\u0634\u0627\u0646 \u062f\u0627\u062f\u0647 \u0645\u06cc \u0634\u0648\u062f.<\/p>\n<pre><code class=\"hljs\">embedding_matrix = zeros((vocab_size, <span class=\"hljs-number\">100<\/span>))\n<span class=\"hljs-keyword\">for<\/span> word, index <span class=\"hljs-keyword\">in<\/span> tokenizer.word_index.items():\n    embedding_vector = embeddings_dictionary.get(word)\n    <span class=\"hljs-keyword\">if<\/span> embedding_vector <span class=\"hljs-keyword\">is<\/span> <span class=\"hljs-keyword\">not<\/span> <span class=\"hljs-literal\">None<\/span>:\n        embedding_matrix(index) = embedding_vector\n<\/code><\/pre>\n<p>\u067e\u0633 \u0627\u0632 \u0627\u062a\u0645\u0627\u0645 \u0645\u0631\u062d\u0644\u0647 embedding \u06a9\u0644\u0645\u0647\u060c \u0645\u0627 \u0622\u0645\u0627\u062f\u0647 \u0627\u06cc\u062c\u0627\u062f \u0645\u062f\u0644 \u062e\u0648\u062f \u0647\u0633\u062a\u06cc\u0645.  \u0645\u0627 \u0627\u0632 API \u0639\u0645\u0644\u06a9\u0631\u062f\u06cc Keras \u0628\u0631\u0627\u06cc \u0627\u06cc\u062c\u0627\u062f \u0645\u062f\u0644 \u062e\u0648\u062f \u0627\u0633\u062a\u0641\u0627\u062f\u0647 \u062e\u0648\u0627\u0647\u06cc\u0645 \u06a9\u0631\u062f.  \u0627\u06af\u0631\u0686\u0647 \u0645\u062f\u0644\u200c\u0647\u0627\u06cc \u0648\u0631\u0648\u062f\u06cc \u062a\u06a9\u06cc \u0645\u0627\u0646\u0646\u062f \u0645\u062f\u0644\u06cc \u06a9\u0647 \u0627\u06a9\u0646\u0648\u0646 \u0627\u06cc\u062c\u0627\u062f \u0645\u06cc\u200c\u06a9\u0646\u06cc\u0645 \u0631\u0627 \u0645\u06cc\u200c\u062a\u0648\u0627\u0646 \u0628\u0627 \u0627\u0633\u062a\u0641\u0627\u062f\u0647 \u0627\u0632 API \u0645\u062a\u0648\u0627\u0644\u06cc \u0646\u06cc\u0632 \u062a\u0648\u0633\u0639\u0647 \u062f\u0627\u062f\u060c \u0627\u0645\u0627 \u0647\u0645\u0627\u0646\u0637\u0648\u0631 \u06a9\u0647 \u062f\u0631 \u0628\u062e\u0634 \u0628\u0639\u062f\u06cc \u0645\u06cc\u200c\u062e\u0648\u0627\u0647\u06cc\u0645 \u06cc\u06a9 \u0645\u062f\u0644 \u0648\u0631\u0648\u062f\u06cc \u0686\u0646\u062f\u06af\u0627\u0646\u0647 \u0631\u0627 \u062a\u0648\u0633\u0639\u0647 \u062f\u0647\u06cc\u0645 \u06a9\u0647 \u0641\u0642\u0637 \u0628\u0627 \u0627\u0633\u062a\u0641\u0627\u062f\u0647 \u0627\u0632 API \u0639\u0645\u0644\u06a9\u0631\u062f\u06cc Keras \u0642\u0627\u0628\u0644 \u062a\u0648\u0633\u0639\u0647 \u0627\u0633\u062a\u060c \u0645\u0627 \u0628\u0647 \u062a\u0627\u0628\u0639\u06cc \u0645\u06cc\u200c\u0645\u0627\u0646\u06cc\u0645. API \u062f\u0631 \u0627\u06cc\u0646 \u0628\u062e\u0634 \u0646\u06cc\u0632 \u0648\u062c\u0648\u062f \u062f\u0627\u0631\u062f.<\/p>\n<p>\u0645\u0627 \u06cc\u06a9 \u0645\u062f\u0644 \u0628\u0633\u06cc\u0627\u0631 \u0633\u0627\u062f\u0647 \u0628\u0627 \u06cc\u06a9 \u0644\u0627\u06cc\u0647 \u0648\u0631\u0648\u062f\u06cc (\u0644\u0627\u06cc\u0647 \u062c\u0627\u0633\u0627\u0632\u06cc)\u060c \u06cc\u06a9 \u0644\u0627\u06cc\u0647 LSTM \u0628\u0627 128 \u0646\u0648\u0631\u0648\u0646 \u0648 \u06cc\u06a9 \u0644\u0627\u06cc\u0647 \u0645\u062a\u0631\u0627\u06a9\u0645 \u0627\u06cc\u062c\u0627\u062f \u062e\u0648\u0627\u0647\u06cc\u0645 \u06a9\u0631\u062f \u06a9\u0647 \u0628\u0647 \u0639\u0646\u0648\u0627\u0646 \u0644\u0627\u06cc\u0647 \u062e\u0631\u0648\u062c\u06cc \u0646\u06cc\u0632 \u0639\u0645\u0644 \u062e\u0648\u0627\u0647\u062f \u06a9\u0631\u062f.  \u0627\u0632 \u0622\u0646\u062c\u0627\u06cc\u06cc \u06a9\u0647 \u0645\u0627 3 \u062e\u0631\u0648\u062c\u06cc \u0645\u0645\u06a9\u0646 \u062f\u0627\u0631\u06cc\u0645\u060c \u062a\u0639\u062f\u0627\u062f \u0646\u0648\u0631\u0648\u0646 \u0647\u0627 3 \u0648 \u062a\u0627\u0628\u0639 \u0641\u0639\u0627\u0644 \u0633\u0627\u0632\u06cc \u062e\u0648\u0627\u0647\u062f \u0628\u0648\u062f <code>softmax<\/code>.  \u0645\u0627 \u0627\u0633\u062a\u0641\u0627\u062f\u0647 \u062e\u0648\u0627\u0647\u06cc\u0645 \u06a9\u0631\u062f <code>categorical_crossentropy<\/code> \u0628\u0647 \u0639\u0646\u0648\u0627\u0646 \u062a\u0627\u0628\u0639 \u0636\u0631\u0631 \u0645\u0627 \u0648 <code>adam<\/code> \u0628\u0647 \u0639\u0646\u0648\u0627\u0646 \u062a\u0627\u0628\u0639 \u0628\u0647\u06cc\u0646\u0647 \u0633\u0627\u0632\u06cc<\/p>\n<pre><code class=\"hljs\">deep_inputs = Input(shape=(maxlen,))\nembedding_layer = Embedding(vocab_size, <span class=\"hljs-number\">100<\/span>, weights=(embedding_matrix), trainable=<span class=\"hljs-literal\">False<\/span>)(deep_inputs)\nLSTM_Layer_1 = LSTM(<span class=\"hljs-number\">128<\/span>)(embedding_layer)\ndense_layer_1 = Dense(<span class=\"hljs-number\">3<\/span>, activation=<span class=\"hljs-string\">'softmax'<\/span>)(LSTM_Layer_1)\nmodel = Model(inputs=deep_inputs, outputs=dense_layer_1)\n\nmodel.<span class=\"hljs-built_in\">compile<\/span>(loss=<span class=\"hljs-string\">'categorical_crossentropy'<\/span>, optimizer=<span class=\"hljs-string\">'adam'<\/span>, metrics=(<span class=\"hljs-string\">'acc'<\/span>))\n<\/code><\/pre>\n<p>\u0627\u062c\u0627\u0632\u0647 \u062f\u0647\u06cc\u062f print \u062e\u0644\u0627\u0635\u0647 \u0645\u062f\u0644 \u0645\u0627:<\/p>\n<pre><code class=\"hljs\"><span class=\"hljs-built_in\">print<\/span>(model.summary())\n<\/code><\/pre>\n<pre><code class=\"hljs\">_________________________________________________________________\nLayer (type)                 Output Shape              Param #\n=================================================================\ninput_1 (InputLayer)         (None, 200)               0\n_________________________________________________________________\nembedding_1 (Embedding)      (None, 200, 100)          5572900\n_________________________________________________________________\nlstm_1 (LSTM)                (None, 128)               117248\n_________________________________________________________________\ndense_1 (Dense)              (None, 3)                 387\n=================================================================\nTotal params: 5,690,535\nTrainable params: 117,635\nNon-trainable params: 5,572,900\n<\/code><\/pre>\n<p>\u062f\u0631 \u0646\u0647\u0627\u06cc\u062a\u060c \u0627\u062c\u0627\u0632\u0647 \u0645\u06cc \u062f\u0647\u062f print \u0628\u0644\u0648\u06a9 \u062f\u06cc\u0627\u06af\u0631\u0627\u0645 \u0634\u0628\u06a9\u0647 \u0639\u0635\u0628\u06cc \u0645\u0627:<\/p>\n<pre><code class=\"hljs\"><span class=\"hljs-keyword\">from<\/span> keras.utils <span class=\"hljs-keyword\">import<\/span> plot_model\nplot_model(model, to_file=<span class=\"hljs-string\">'model_plot1.png'<\/span>, show_shapes=<span class=\"hljs-literal\">True<\/span>, show_layer_names=<span class=\"hljs-literal\">True<\/span>)\n<\/code><\/pre>\n<p>\u067e\u0631\u0648\u0646\u062f\u0647 <code>model_plot1.png<\/code> \u062f\u0631 \u0645\u0633\u06cc\u0631 \u0641\u0627\u06cc\u0644 \u0645\u062d\u0644\u06cc \u0634\u0645\u0627 \u0627\u06cc\u062c\u0627\u062f \u062e\u0648\u0627\u0647\u062f \u0634\u062f.  \u0627\u06af\u0631 \u062a\u0635\u0648\u06cc\u0631 \u0631\u0627 \u0628\u0627\u0632 \u06a9\u0646\u06cc\u062f \u0628\u0647 \u0634\u06a9\u0644 \u0632\u06cc\u0631 \u062f\u0631 \u0645\u06cc \u0622\u06cc\u062f:<\/p>\n<p><img decoding=\"async\" class=\"img-responsive\" src=\"https:\/\/rasanegar.com\/blog\/wp-content\/uploads\/2024\/01\/python-for-nlp-creating-text-classification-model-with-multiple-inputs-in-keras-3.png\" alt=\"\u0633\u0631\" title=\"\"><\/p>\n<p>\u0645\u06cc \u0628\u06cc\u0646\u06cc\u062f \u06a9\u0647 \u0645\u062f\u0644 \u062f\u0627\u0631\u0627\u06cc 1 \u0644\u0627\u06cc\u0647 \u0648\u0631\u0648\u062f\u06cc\u060c 1 \u0644\u0627\u06cc\u0647 \u062a\u0639\u0628\u06cc\u0647 \u0634\u062f\u0647\u060c 1 LSTM \u0648 \u06cc\u06a9 \u0644\u0627\u06cc\u0647 \u0645\u062a\u0631\u0627\u06a9\u0645 \u0627\u0633\u062a \u06a9\u0647 \u0628\u0647 \u0639\u0646\u0648\u0627\u0646 \u0644\u0627\u06cc\u0647 \u062e\u0631\u0648\u062c\u06cc \u0646\u06cc\u0632 \u0639\u0645\u0644 \u0645\u06cc \u06a9\u0646\u062f.<\/p>\n<p>\u062d\u0627\u0644 \u0628\u06cc\u0627\u06cc\u06cc\u062f \u0645\u062f\u0644 \u062e\u0648\u062f \u0631\u0627 \u0622\u0645\u0648\u0632\u0634 \u062f\u0647\u06cc\u0645:<\/p>\n<pre><code class=\"hljs\">history = model.fit(X_train, y_train, batch_size=<span class=\"hljs-number\">128<\/span>, epochs=<span class=\"hljs-number\">10<\/span>, verbose=<span class=\"hljs-number\">1<\/span>, validation_split=<span class=\"hljs-number\">0.2<\/span>)\n<\/code><\/pre>\n<p>\u0645\u062f\u0644 \u0622\u0645\u0648\u0632\u0634 \u062f\u0627\u062f\u0647 \u062e\u0648\u0627\u0647\u062f \u0634\u062f \u0631\u0648\u06cc 80 \u062f\u0631\u0635\u062f \u0627\u0632 \u062f\u0627\u062f\u0647 \u0647\u0627\u06cc \u0642\u0637\u0627\u0631 \u0648 \u0627\u0639\u062a\u0628\u0627\u0631\u0633\u0646\u062c\u06cc \u062e\u0648\u0627\u0647\u062f \u0634\u062f \u0631\u0648\u06cc 20 \u062f\u0631\u0635\u062f \u0627\u0632 \u062f\u0627\u062f\u0647 \u0647\u0627\u06cc \u0642\u0637\u0627\u0631.  \u0646\u062a\u0627\u06cc\u062c \u0628\u0631\u0627\u06cc 10 \u062f\u0648\u0631\u0647 \u0628\u0647 \u0634\u0631\u062d \u0632\u06cc\u0631 \u0627\u0633\u062a:<\/p>\n<pre><code class=\"hljs\">Train \u0631\u0648\u06cc 32000 samples, validate \u0631\u0648\u06cc 8000 samples\nEpoch 1\/10\n32000\/32000 (==============================) - 81s 3ms\/step - loss: 0.8640 - acc: 0.6623 - val_loss: 0.8356 - val_acc: 0.6730\nEpoch 2\/10\n32000\/32000 (==============================) - 80s 3ms\/step - loss: 0.8508 - acc: 0.6618 - val_loss: 0.8399 - val_acc: 0.6690\nEpoch 3\/10\n32000\/32000 (==============================) - 84s 3ms\/step - loss: 0.8461 - acc: 0.6647 - val_loss: 0.8374 - val_acc: 0.6726\nEpoch 4\/10\n32000\/32000 (==============================) - 82s 3ms\/step - loss: 0.8288 - acc: 0.6709 - val_loss: 0.7392 - val_acc: 0.6861\nEpoch 5\/10\n32000\/32000 (==============================) - 82s 3ms\/step - loss: 0.7444 - acc: 0.6804 - val_loss: 0.6371 - val_acc: 0.7311\nEpoch 6\/10\n32000\/32000 (==============================) - 83s 3ms\/step - loss: 0.5969 - acc: 0.7484 - val_loss: 0.5602 - val_acc: 0.7682\nEpoch 7\/10\n32000\/32000 (==============================) - 82s 3ms\/step - loss: 0.5484 - acc: 0.7623 - val_loss: 0.5244 - val_acc: 0.7814\nEpoch 8\/10\n32000\/32000 (==============================) - 86s 3ms\/step - loss: 0.5052 - acc: 0.7866 - val_loss: 0.4971 - val_acc: 0.7950\nEpoch 9\/10\n32000\/32000 (==============================) - 84s 3ms\/step - loss: 0.4753 - acc: 0.8032 - val_loss: 0.4839 - val_acc: 0.7965\nEpoch 10\/10\n32000\/32000 (==============================) - 82s 3ms\/step - loss: 0.4539 - acc: 0.8110 - val_loss: 0.4622 - val_acc: 0.8046\n<\/code><\/pre>\n<p>\u0645\u06cc \u0628\u06cc\u0646\u06cc\u062f \u06a9\u0647 \u062f\u0642\u062a \u0622\u0645\u0648\u0632\u0634\u06cc \u0646\u0647\u0627\u06cc\u06cc \u0645\u062f\u0644 81.10 \u062f\u0631\u0635\u062f \u0627\u0633\u062a \u062f\u0631 \u062d\u0627\u0644\u06cc \u06a9\u0647 \u062f\u0642\u062a \u0627\u0639\u062a\u0628\u0627\u0631\u0633\u0646\u062c\u06cc 80.46 \u0627\u0633\u062a.  \u062a\u0641\u0627\u0648\u062a \u0628\u0633\u06cc\u0627\u0631 \u06a9\u0648\u0686\u06a9 \u0627\u0633\u062a \u0648 \u0628\u0646\u0627\u0628\u0631\u0627\u06cc\u0646 \u0645\u0627 \u0641\u0631\u0636 \u0645\u06cc \u06a9\u0646\u06cc\u0645 \u06a9\u0647 \u0645\u062f\u0644 \u0645\u0627 \u0628\u06cc\u0634 \u0627\u0632 \u062d\u062f \u0645\u0646\u0627\u0633\u0628 \u0646\u06cc\u0633\u062a \u0631\u0648\u06cc \u062f\u0627\u062f\u0647 \u0647\u0627\u06cc \u0622\u0645\u0648\u0632\u0634\u06cc<\/p>\n<p>\u062d\u0627\u0644 \u0628\u06cc\u0627\u06cc\u06cc\u062f \u0639\u0645\u0644\u06a9\u0631\u062f \u0645\u062f\u0644 \u062e\u0648\u062f \u0631\u0627 \u0627\u0631\u0632\u06cc\u0627\u0628\u06cc \u06a9\u0646\u06cc\u0645 \u0631\u0648\u06cc \u0645\u062c\u0645\u0648\u0639\u0647 \u062a\u0633\u062a:<\/p>\n<pre><code class=\"hljs\">score = model.evaluate(X_test, y_test, verbose=<span class=\"hljs-number\">1<\/span>)\n\n<span class=\"hljs-built_in\">print<\/span>(<span class=\"hljs-string\">\"Test Score:\"<\/span>, score(<span class=\"hljs-number\">0<\/span>))\n<span class=\"hljs-built_in\">print<\/span>(<span class=\"hljs-string\">\"Test Accuracy:\"<\/span>, score(<span class=\"hljs-number\">1<\/span>))\n<\/code><\/pre>\n<p>\u062e\u0631\u0648\u062c\u06cc \u0628\u0647 \u0634\u06a9\u0644 \u0632\u06cc\u0631 \u0627\u0633\u062a:<\/p>\n<pre><code class=\"hljs\">10000\/10000 (==============================) - 37s 4ms\/step\nTest Score: 0.4592904740810394\nTest Accuracy: 0.8101\n<\/code><\/pre>\n<p>\u062f\u0631 \u0646\u0647\u0627\u06cc\u062a\u060c \u0628\u06cc\u0627\u06cc\u06cc\u062f \u0645\u0642\u0627\u062f\u06cc\u0631 \u0627\u0632 \u062f\u0633\u062a \u062f\u0627\u062f\u0646 \u0648 \u062f\u0642\u062a \u0631\u0627 \u0628\u0631\u0627\u06cc \u0645\u062c\u0645\u0648\u0639\u0647 \u0647\u0627\u06cc \u0622\u0645\u0648\u0632\u0634\u06cc \u0648 \u0622\u0632\u0645\u0627\u06cc\u0634\u06cc \u062a\u0631\u0633\u06cc\u0645 \u06a9\u0646\u06cc\u0645:<\/p>\n<pre><code class=\"hljs\"><span class=\"hljs-keyword\">import<\/span> matplotlib.pyplot <span class=\"hljs-keyword\">as<\/span> plt\n\nplt.plot(history.history(<span class=\"hljs-string\">'acc'<\/span>))\nplt.plot(history.history(<span class=\"hljs-string\">'val_acc'<\/span>))\n\nplt.title(<span class=\"hljs-string\">'model accuracy'<\/span>)\nplt.ylabel(<span class=\"hljs-string\">'accuracy'<\/span>)\nplt.xlabel(<span class=\"hljs-string\">'epoch'<\/span>)\nplt.legend((<span class=\"hljs-string\">'train'<\/span>,<span class=\"hljs-string\">'test'<\/span>), loc=<span class=\"hljs-string\">'upper left'<\/span>)\nplt.show()\n\nplt.plot(history.history(<span class=\"hljs-string\">'loss'<\/span>))\nplt.plot(history.history(<span class=\"hljs-string\">'val_loss'<\/span>))\n\nplt.title(<span class=\"hljs-string\">'model loss'<\/span>)\nplt.ylabel(<span class=\"hljs-string\">'loss'<\/span>)\nplt.xlabel(<span class=\"hljs-string\">'epoch'<\/span>)\nplt.legend((<span class=\"hljs-string\">'train'<\/span>,<span class=\"hljs-string\">'test'<\/span>), loc=<span class=\"hljs-string\">'upper left'<\/span>)\nplt.show()\n<\/code><\/pre>\n<p>\u0634\u0645\u0627 \u0628\u0627\u06cc\u062f \u062f\u0648 \u0646\u0645\u0648\u062f\u0627\u0631 \u0632\u06cc\u0631 \u0631\u0627 \u0628\u0628\u06cc\u0646\u06cc\u062f:<\/p>\n<p><img decoding=\"async\" class=\"img-responsive\" src=\"https:\/\/rasanegar.com\/blog\/wp-content\/uploads\/2024\/01\/python-for-nlp-creating-text-classification-model-with-multiple-inputs-in-keras-4.png\" alt=\"\u0633\u0631\" title=\"\"><\/p>\n<p>\u0634\u0645\u0627 \u0645\u06cc \u062a\u0648\u0627\u0646\u06cc\u062f \u062e\u0637\u0648\u0637 \u0631\u0627 \u0628\u0631\u0627\u06cc \u062f\u0642\u062a \u0622\u0645\u0648\u0632\u0634 \u0648 \u062a\u0633\u062a \u0648 \u062a\u0644\u0641\u0627\u062a \u0628\u0633\u06cc\u0627\u0631 \u0646\u0632\u062f\u06cc\u06a9 \u0628\u0647 \u06cc\u06a9\u062f\u06cc\u06af\u0631 \u0645\u0634\u0627\u0647\u062f\u0647 \u06a9\u0646\u06cc\u062f \u06a9\u0647 \u0628\u0647 \u0627\u06cc\u0646 \u0645\u0639\u0646\u06cc \u0627\u0633\u062a \u06a9\u0647 \u0645\u062f\u0644 \u0628\u06cc\u0634 \u0627\u0632 \u062d\u062f \u0645\u0646\u0627\u0633\u0628 \u0646\u06cc\u0633\u062a.<\/p>\n<p>\u062f\u0631 \u0627\u06cc\u0646 \u0628\u062e\u0634\u060c \u06cc\u06a9 \u0645\u062f\u0644 \u0637\u0628\u0642\u0647 \u0628\u0646\u062f\u06cc \u0627\u06cc\u062c\u0627\u062f \u0645\u06cc \u06a9\u0646\u06cc\u0645 \u06a9\u0647 \u0627\u0632 \u0627\u0637\u0644\u0627\u0639\u0627\u062a \u0645\u0648\u062c\u0648\u062f \u062f\u0631 \u0622\u0646 \u0627\u0633\u062a\u0641\u0627\u062f\u0647 \u0645\u06cc \u06a9\u0646\u062f <code>useful<\/code>\u060c <code>funny<\/code>\u060c \u0648 <code>cool<\/code> \u0633\u062a\u0648\u0646 \u0647\u0627\u06cc \u0628\u0631\u0631\u0633\u06cc yelp.  \u0627\u0632 \u0622\u0646\u062c\u0627\u06cc\u06cc \u06a9\u0647 \u062f\u0627\u062f\u0647\u200c\u0647\u0627\u06cc \u0627\u06cc\u0646 \u0633\u062a\u0648\u0646\u200c\u0647\u0627 \u0628\u0647 \u062e\u0648\u0628\u06cc \u0633\u0627\u062e\u062a\u0627\u0631 \u06cc\u0627\u0641\u062a\u0647\u200c\u0627\u0646\u062f \u0648 \u062d\u0627\u0648\u06cc \u0647\u06cc\u0686 \u0627\u0644\u06af\u0648\u06cc \u062a\u0631\u062a\u06cc\u0628\u06cc \u06cc\u0627 \u0641\u0636\u0627\u06cc\u06cc \u0646\u06cc\u0633\u062a\u0646\u062f\u060c \u0645\u06cc\u200c\u062a\u0648\u0627\u0646\u06cc\u0645 \u0627\u0632 \u0634\u0628\u06a9\u0647\u200c\u0647\u0627\u06cc \u0639\u0635\u0628\u06cc \u0645\u062a\u0635\u0644 \u0645\u062a\u0631\u0627\u06a9\u0645 \u0633\u0627\u062f\u0647 \u0628\u0631\u0627\u06cc \u067e\u06cc\u0634\u200c\u0628\u06cc\u0646\u06cc \u0627\u0633\u062a\u0641\u0627\u062f\u0647 \u06a9\u0646\u06cc\u0645.<\/p>\n<p>\u0628\u06cc\u0627\u06cc\u06cc\u062f \u0645\u06cc\u0627\u0646\u06af\u06cc\u0646 \u0634\u0645\u0627\u0631\u0634 \u0647\u0627 \u0631\u0627 \u062a\u0631\u0633\u06cc\u0645 \u06a9\u0646\u06cc\u0645 <code>useful<\/code>\u060c <code>funny<\/code>\u060c \u0648 <code>cool<\/code> \u0628\u0631\u0631\u0633\u06cc \u0647\u0627 \u062f\u0631 \u0645\u0642\u0627\u0628\u0644 \u0646\u0645\u0631\u0647 \u0628\u0627\u0632\u0628\u06cc\u0646\u06cc<\/p>\n<pre><code class=\"hljs\"><span class=\"hljs-keyword\">import<\/span> seaborn <span class=\"hljs-keyword\">as<\/span> sns\nsns.barplot(x=<span class=\"hljs-string\">'reviews_score'<\/span>, y=<span class=\"hljs-string\">'useful'<\/span>, data=yelp_reviews)\n<\/code><\/pre>\n<p><img decoding=\"async\" class=\"img-responsive\" src=\"https:\/\/rasanegar.com\/blog\/wp-content\/uploads\/2024\/01\/python-for-nlp-creating-text-classification-model-with-multiple-inputs-in-keras-5.png\" alt=\"\u0633\u0631\" title=\"\"><\/p>\n<p>\u0627\u0632 \u062e\u0631\u0648\u062c\u06cc\u060c \u0645\u06cc\u200c\u062a\u0648\u0627\u0646\u06cc\u062f \u0628\u0628\u06cc\u0646\u06cc\u062f \u06a9\u0647 \u0645\u06cc\u0627\u0646\u06af\u06cc\u0646 \u062a\u0639\u062f\u0627\u062f \u0646\u0638\u0631\u0627\u062a \u0628\u0647\u200c\u0639\u0646\u0648\u0627\u0646 \u0639\u0644\u0627\u0645\u062a\u200c\u06af\u0630\u0627\u0631\u06cc \u0634\u062f\u0647 \u0627\u0633\u062a <code>useful<\/code> \u0628\u0631\u0627\u06cc \u0646\u0642\u062f\u0647\u0627\u06cc \u0628\u062f\u060c \u0628\u0627\u0644\u0627\u062a\u0631\u06cc\u0646 \u0631\u062a\u0628\u0647 \u0627\u0633\u062a \u0648 \u067e\u0633 \u0627\u0632 \u0622\u0646 \u0645\u06cc\u0627\u0646\u06af\u06cc\u0646 \u0646\u0638\u0631\u0627\u062a \u0648 \u0646\u0642\u062f\u0647\u0627\u06cc \u062e\u0648\u0628 \u0642\u0631\u0627\u0631 \u062f\u0627\u0631\u0646\u062f.<\/p>\n<p>\u062d\u0627\u0644\u0627 \u0628\u06cc\u0627\u06cc\u06cc\u062f \u0645\u06cc\u0627\u0646\u06af\u06cc\u0646 \u0634\u0645\u0627\u0631\u0634 \u0631\u0627 \u0631\u0633\u0645 \u06a9\u0646\u06cc\u0645 <code>funny<\/code> \u0628\u0631\u0631\u0633\u06cc \u0647\u0627:<\/p>\n<pre><code class=\"hljs\">sns.barplot(x=<span class=\"hljs-string\">'reviews_score'<\/span>, y=<span class=\"hljs-string\">'funny'<\/span>, data=yelp_reviews)\n<\/code><\/pre>\n<p><img decoding=\"async\" class=\"img-responsive\" src=\"https:\/\/rasanegar.com\/blog\/wp-content\/uploads\/2024\/01\/python-for-nlp-creating-text-classification-model-with-multiple-inputs-in-keras-6.png\" alt=\"\u0633\u0631\" title=\"\"><\/p>\n<p>\u062e\u0631\u0648\u062c\u06cc \u0646\u0634\u0627\u0646 \u0645\u06cc \u062f\u0647\u062f \u06a9\u0647 \u0645\u062c\u062f\u062f\u0627\u064b\u060c \u0645\u06cc\u0627\u0646\u06af\u06cc\u0646 \u062a\u0639\u062f\u0627\u062f \u0646\u0638\u0631\u0627\u062a \u0628\u0647 \u0639\u0646\u0648\u0627\u0646 \u0639\u0644\u0627\u0645\u062a \u06af\u0630\u0627\u0631\u06cc \u0634\u062f\u0647 \u0627\u0633\u062a <code>funny<\/code> \u0628\u0627\u0644\u0627\u062a\u0631\u06cc\u0646 \u0628\u0631\u0627\u06cc \u0628\u0631\u0631\u0633\u06cc \u0647\u0627\u06cc \u0628\u062f \u0627\u0633\u062a.<\/p>\n<p>\u062f\u0631 \u0646\u0647\u0627\u06cc\u062a\u060c \u0628\u06cc\u0627\u06cc\u06cc\u062f \u0645\u0642\u062f\u0627\u0631 \u0645\u06cc\u0627\u0646\u06af\u06cc\u0646 \u0631\u0627 \u062a\u0631\u0633\u06cc\u0645 \u06a9\u0646\u06cc\u0645 <code>cool<\/code> \u0633\u062a\u0648\u0646 \u0645\u0642\u0627\u0628\u0644 <code>reviews_score<\/code> \u0633\u062a\u0648\u0646  \u0645\u0627 \u0627\u0646\u062a\u0638\u0627\u0631 \u062f\u0627\u0631\u06cc\u0645 \u06a9\u0647 \u0645\u06cc\u0627\u0646\u06af\u06cc\u0646 \u0634\u0645\u0627\u0631\u0634 \u0628\u0631\u0627\u06cc <code>cool<\/code> \u0633\u062a\u0648\u0646 \u0628\u0631\u0627\u06cc \u0628\u0631\u0631\u0633\u06cc \u0647\u0627\u06cc \u062e\u0648\u0628 \u0628\u0627\u0644\u0627\u062a\u0631\u06cc\u0646 \u062e\u0648\u0627\u0647\u062f \u0628\u0648\u062f \u0632\u06cc\u0631\u0627 \u0645\u0631\u062f\u0645 \u0627\u063a\u0644\u0628 \u0646\u0638\u0631\u0627\u062a \u0645\u062b\u0628\u062a \u06cc\u0627 \u062e\u0648\u0628 \u0631\u0627 \u0628\u0647 \u0639\u0646\u0648\u0627\u0646 \u062c\u0627\u0644\u0628 \u0639\u0644\u0627\u0645\u062a \u06af\u0630\u0627\u0631\u06cc \u0645\u06cc \u06a9\u0646\u0646\u062f:<\/p>\n<pre><code class=\"hljs\">sns.barplot(x=<span class=\"hljs-string\">'reviews_score'<\/span>, y=<span class=\"hljs-string\">'cool'<\/span>, data=yelp_reviews)\n<\/code><\/pre>\n<p><img decoding=\"async\" class=\"img-responsive\" src=\"https:\/\/rasanegar.com\/blog\/wp-content\/uploads\/2024\/01\/python-for-nlp-creating-text-classification-model-with-multiple-inputs-in-keras-7.png\" alt=\"\u0633\u0631\" title=\"\"><\/p>\n<p>\u0647\u0645\u0627\u0646\u0637\u0648\u0631 \u06a9\u0647 \u0627\u0646\u062a\u0638\u0627\u0631 \u0645\u06cc \u0631\u0648\u062f\u060c \u0645\u06cc\u0627\u0646\u06af\u06cc\u0646 <code>cool<\/code> \u0634\u0645\u0627\u0631\u0634 \u0628\u0631\u0627\u06cc \u0628\u0631\u0631\u0633\u06cc \u062e\u0648\u0628 \u0628\u0627\u0644\u0627\u062a\u0631\u06cc\u0646 \u0627\u0633\u062a.  \u0627\u0632 \u0627\u06cc\u0646 \u0627\u0637\u0644\u0627\u0639\u0627\u062a\u060c \u0645\u06cc \u062a\u0648\u0627\u0646\u06cc\u0645 \u0628\u0627 \u062e\u06cc\u0627\u0644 \u0631\u0627\u062d\u062a \u0641\u0631\u0636 \u06a9\u0646\u06cc\u0645 \u06a9\u0647 \u062a\u0639\u062f\u0627\u062f \u0645\u0642\u0627\u062f\u06cc\u0631 \u0628\u0631\u0627\u06cc <code>useful<\/code>\u060c <code>funny<\/code>\u060c \u0648 <code>cool<\/code> \u0633\u062a\u0648\u0646 \u0647\u0627 \u062a\u0627 \u062d\u062f\u0648\u062f\u06cc \u0628\u0627 <code>reviews_score<\/code> \u0633\u062a\u0648\u0646 \u0647\u0627.  \u0628\u0646\u0627\u0628\u0631\u0627\u06cc\u0646\u060c \u0645\u0627 \u0633\u0639\u06cc \u062e\u0648\u0627\u0647\u06cc\u0645 \u06a9\u0631\u062f \u0627\u0632 \u062f\u0627\u062f\u0647 \u0647\u0627\u06cc \u0627\u06cc\u0646 \u0633\u0647 \u0633\u062a\u0648\u0646 \u0628\u0631\u0627\u06cc \u0622\u0645\u0648\u0632\u0634 \u0627\u0644\u06af\u0648\u0631\u06cc\u062a\u0645 \u062e\u0648\u062f \u0627\u0633\u062a\u0641\u0627\u062f\u0647 \u06a9\u0646\u06cc\u0645 \u06a9\u0647 \u0645\u0642\u062f\u0627\u0631 \u0631\u0627 \u0628\u0631\u0627\u06cc <code>reviews_score<\/code> \u0633\u062a\u0648\u0646<\/p>\n<p>\u0628\u06cc\u0627\u06cc\u06cc\u062f \u0627\u06cc\u0646 \u0633\u0647 \u0633\u062a\u0648\u0646 \u0631\u0627 \u0627\u0632 \u0645\u062c\u0645\u0648\u0639\u0647 \u062f\u0627\u062f\u0647 \u062e\u0648\u062f \u0641\u06cc\u0644\u062a\u0631 \u06a9\u0646\u06cc\u0645:<\/p>\n<pre><code class=\"hljs\">yelp_reviews_meta = yelp_reviews((<span class=\"hljs-string\">'useful'<\/span>, <span class=\"hljs-string\">'funny'<\/span>, <span class=\"hljs-string\">'cool'<\/span>))\n\nX = yelp_reviews_meta.values\n\ny = yelp_reviews(<span class=\"hljs-string\">'reviews_score'<\/span>)\n<\/code><\/pre>\n<p>\u062f\u0631 \u0645\u0631\u062d\u0644\u0647 \u0628\u0639\u062f\u060c \u0645\u0627 \u0628\u0631\u0686\u0633\u0628\u200c\u0647\u0627\u06cc \u062e\u0648\u062f \u0631\u0627 \u0628\u0647 \u0645\u0642\u0627\u062f\u06cc\u0631 \u0631\u0645\u0632\u06af\u0630\u0627\u0631\u06cc \u0634\u062f\u0647 \u062a\u0628\u062f\u06cc\u0644 \u0645\u06cc\u200c\u06a9\u0646\u06cc\u0645 \u0648 \u0633\u067e\u0633 \u062f\u0627\u062f\u0647\u200c\u0647\u0627\u06cc \u062e\u0648\u062f \u0631\u0627 \u0628\u0647 \u0645\u062c\u0645\u0648\u0639\u0647\u200c\u0647\u0627\u06cc \u0642\u0637\u0627\u0631 \u0648 \u0622\u0632\u0645\u0627\u06cc\u0634 \u062a\u0642\u0633\u06cc\u0645 \u0645\u06cc\u200c\u06a9\u0646\u06cc\u0645:<\/p>\n<pre><code class=\"hljs\"><span class=\"hljs-keyword\">from<\/span> sklearn <span class=\"hljs-keyword\">import<\/span> preprocessing\n\n\nlabel_encoder = preprocessing.LabelEncoder()\n\n\ny = label_encoder.fit_transform(y)\n\nX_train, X_test, y_train, y_test = train_test_split(X, y, test_size=<span class=\"hljs-number\">0.20<\/span>, random_state=<span class=\"hljs-number\">42<\/span>)\n\n<span class=\"hljs-keyword\">from<\/span> keras.utils <span class=\"hljs-keyword\">import<\/span> to_categorical\ny_train = to_categorical(y_train)\ny_test = to_categorical(y_test)\n<\/code><\/pre>\n<p>\u0645\u0631\u062d\u0644\u0647 \u0628\u0639\u062f\u06cc \u0627\u06cc\u062c\u0627\u062f \u0645\u062f\u0644 \u0645\u0627 \u0627\u0633\u062a.  \u0645\u062f\u0644 \u0645\u0627 \u0627\u0632 \u0686\u0647\u0627\u0631 \u0644\u0627\u06cc\u0647 \u062a\u0634\u06a9\u06cc\u0644 \u0634\u062f\u0647 \u0627\u0633\u062a (\u0634\u0645\u0627 \u0645\u06cc \u062a\u0648\u0627\u0646\u06cc\u062f \u0647\u0631 \u0639\u062f\u062f\u06cc \u0631\u0627 \u0627\u0645\u062a\u062d\u0627\u0646 \u06a9\u0646\u06cc\u062f): \u0644\u0627\u06cc\u0647 \u0648\u0631\u0648\u062f\u06cc\u060c \u062f\u0648 \u0644\u0627\u06cc\u0647 \u067e\u0646\u0647\u0627\u0646 \u0645\u062a\u0631\u0627\u06a9\u0645 \u0628\u0627 10 \u0646\u0648\u0631\u0648\u0646 \u0648 \u0639\u0645\u0644\u06a9\u0631\u062f\u0647\u0627\u06cc \u0641\u0639\u0627\u0644 \u0633\u0627\u0632\u06cc ReLU\u060c \u0648 \u062f\u0631 \u0646\u0647\u0627\u06cc\u062a \u06cc\u06a9 \u0644\u0627\u06cc\u0647 \u0645\u062a\u0631\u0627\u06a9\u0645 \u062e\u0631\u0648\u062c\u06cc \u0628\u0627 3 \u0646\u0648\u0631\u0648\u0646 \u0648 \u062a\u0627\u0628\u0639 \u0641\u0639\u0627\u0644 \u0633\u0627\u0632\u06cc softmax.  \u062a\u0627\u0628\u0639 \u0636\u0631\u0631 \u0648 \u0628\u0647\u06cc\u0646\u0647 \u0633\u0627\u0632 \u062e\u0648\u0627\u0647\u062f \u0628\u0648\u062f <code>categorical_crossentropy<\/code> \u0648 <code>adam<\/code>\u060c \u0628\u0647 \u062a\u0631\u062a\u06cc\u0628.<\/p>\n<p>\u0627\u0633\u06a9\u0631\u06cc\u067e\u062a \u0632\u06cc\u0631 \u0645\u062f\u0644 \u0631\u0627 \u062a\u0639\u0631\u06cc\u0641 \u0645\u06cc \u06a9\u0646\u062f:<\/p>\n<pre><code class=\"hljs\">input2 = Input(shape=(<span class=\"hljs-number\">3<\/span>,))\ndense_layer_1 = Dense(<span class=\"hljs-number\">10<\/span>, activation=<span class=\"hljs-string\">'relu'<\/span>)(input2)\ndense_layer_2 = Dense(<span class=\"hljs-number\">10<\/span>, activation=<span class=\"hljs-string\">'relu'<\/span>)(dense_layer_1)\noutput = Dense(<span class=\"hljs-number\">3<\/span>, activation=<span class=\"hljs-string\">'softmax'<\/span>)(dense_layer_2)\n\nmodel = Model(inputs=input2, outputs=output)\nmodel.<span class=\"hljs-built_in\">compile<\/span>(loss=<span class=\"hljs-string\">'categorical_crossentropy'<\/span>, optimizer=<span class=\"hljs-string\">'adam'<\/span>, metrics=(<span class=\"hljs-string\">'acc'<\/span>))\n<\/code><\/pre>\n<p>\u0627\u062c\u0627\u0632\u0647 \u062f\u0647\u06cc\u062f print \u062e\u0644\u0627\u0635\u0647 \u0645\u062f\u0644:<\/p>\n<pre><code class=\"hljs\"><span class=\"hljs-built_in\">print<\/span>(model.summary())\n<\/code><\/pre>\n<pre><code class=\"hljs\">_________________________________________________________________\nLayer (type)                 Output Shape              Param #\n=================================================================\ninput_1 (InputLayer)         (None, 3)                 0\n_________________________________________________________________\ndense_1 (Dense)              (None, 10)                40\n_________________________________________________________________\ndense_2 (Dense)              (None, 10)                110\n_________________________________________________________________\ndense_3 (Dense)              (None, 3)                 33\n=================================================================\nTotal params: 183\nTrainable params: 183\nNon-trainable params: 0\n<\/code><\/pre>\n<p>\u062f\u0631 \u0646\u0647\u0627\u06cc\u062a\u060c \u0628\u0644\u0648\u06a9 \u062f\u06cc\u0627\u06af\u0631\u0627\u0645 \u0628\u0631\u0627\u06cc \u0645\u062f\u0644 \u0631\u0627 \u0645\u06cc \u062a\u0648\u0627\u0646 \u0627\u0632 \u0637\u0631\u06cc\u0642 \u0627\u0633\u06a9\u0631\u06cc\u067e\u062a \u0632\u06cc\u0631 \u0627\u06cc\u062c\u0627\u062f \u06a9\u0631\u062f:<\/p>\n<pre><code class=\"hljs\"><span class=\"hljs-keyword\">from<\/span> keras.utils <span class=\"hljs-keyword\">import<\/span> plot_model\nplot_model(model, to_file=<span class=\"hljs-string\">'model_plot2.png'<\/span>, show_shapes=<span class=\"hljs-literal\">True<\/span>, show_layer_names=<span class=\"hljs-literal\">True<\/span>)\n<\/code><\/pre>\n<p>\u062d\u0627\u0644\u0627 \u0627\u06af\u0631 \u062f\u0631 \u0631\u0627 \u0628\u0627\u0632 \u06a9\u0646\u06cc\u062f <code>model_plot2.png<\/code> \u0641\u0627\u06cc\u0644 \u0627\u0632 \u0645\u0633\u06cc\u0631 \u0641\u0627\u06cc\u0644 \u0645\u062d\u0644\u06cc \u0634\u0645\u0627\u060c \u0628\u0647 \u0635\u0648\u0631\u062a \u0632\u06cc\u0631 \u0627\u0633\u062a:<\/p>\n<p><img decoding=\"async\" class=\"img-responsive\" src=\"https:\/\/rasanegar.com\/blog\/wp-content\/uploads\/2024\/01\/python-for-nlp-creating-text-classification-model-with-multiple-inputs-in-keras-8.png\" alt=\"\u0633\u0631\" title=\"\"><\/p>\n<p>\u062d\u0627\u0644\u0627 \u0628\u06cc\u0627\u06cc\u06cc\u062f \u0645\u062f\u0644 \u0648 print \u0645\u0642\u0627\u062f\u06cc\u0631 \u062f\u0642\u062a \u0648 \u0636\u0631\u0631 \u0628\u0631\u0627\u06cc \u0647\u0631 \u062f\u0648\u0631\u0647:<\/p>\n<pre><code class=\"hljs\">history = model.fit(X_train, y_train, batch_size=<span class=\"hljs-number\">16<\/span>, epochs=<span class=\"hljs-number\">10<\/span>, verbose=<span class=\"hljs-number\">1<\/span>, validation_split=<span class=\"hljs-number\">0.2<\/span>)\n<\/code><\/pre>\n<pre><code class=\"hljs\">Train \u0631\u0648\u06cc 32000 samples, validate \u0631\u0648\u06cc 8000 samples\nEpoch 1\/10\n32000\/32000 (==============================) - 8s 260us\/step - loss: 0.8429 - acc: 0.6649 - val_loss: 0.8166 - val_acc: 0.6734\nEpoch 2\/10\n32000\/32000 (==============================) - 7s 214us\/step - loss: 0.8203 - acc: 0.6685 - val_loss: 0.8156 - val_acc: 0.6737\nEpoch 3\/10\n32000\/32000 (==============================) - 7s 217us\/step - loss: 0.8187 - acc: 0.6685 - val_loss: 0.8150 - val_acc: 0.6736\nEpoch 4\/10\n32000\/32000 (==============================) - 7s 220us\/step - loss: 0.8183 - acc: 0.6695 - val_loss: 0.8160 - val_acc: 0.6740\nEpoch 5\/10\n32000\/32000 (==============================) - 7s 227us\/step - loss: 0.8177 - acc: 0.6686 - val_loss: 0.8149 - val_acc: 0.6751\nEpoch 6\/10\n32000\/32000 (==============================) - 7s 219us\/step - loss: 0.8175 - acc: 0.6686 - val_loss: 0.8157 - val_acc: 0.6744\nEpoch 7\/10\n32000\/32000 (==============================) - 7s 216us\/step - loss: 0.8172 - acc: 0.6696 - val_loss: 0.8145 - val_acc: 0.6733\nEpoch 8\/10\n32000\/32000 (==============================) - 7s 214us\/step - loss: 0.8175 - acc: 0.6689 - val_loss: 0.8139 - val_acc: 0.6734\nEpoch 9\/10\n32000\/32000 (==============================) - 7s 215us\/step - loss: 0.8169 - acc: 0.6691 - val_loss: 0.8160 - val_acc: 0.6744\nEpoch 10\/10\n32000\/32000 (==============================) - 7s 216us\/step - loss: 0.8167 - acc: 0.6694 - val_loss: 0.8138 - val_acc: 0.6736\n<\/code><\/pre>\n<p>\u0627\u0632 \u062e\u0631\u0648\u062c\u06cc\u060c \u0645\u06cc \u062a\u0648\u0627\u0646\u06cc\u062f \u0628\u0628\u06cc\u0646\u06cc\u062f \u06a9\u0647 \u0645\u062f\u0644 \u0645\u0627 \u0647\u0645\u06af\u0631\u0627 \u0646\u06cc\u0633\u062a \u0648 \u0645\u0642\u0627\u062f\u06cc\u0631 \u062f\u0642\u062a \u062f\u0631 \u062a\u0645\u0627\u0645 \u062f\u0648\u0631\u0647 \u0647\u0627 \u0628\u06cc\u0646 66 \u0648 67 \u0628\u0627\u0642\u06cc \u0645\u06cc \u0645\u0627\u0646\u062f.<\/p>\n<p>\u0628\u06cc\u0627\u06cc\u06cc\u062f \u0628\u0628\u06cc\u0646\u06cc\u0645 \u06a9\u0647 \u0645\u062f\u0644 \u0686\u06af\u0648\u0646\u0647 \u0639\u0645\u0644 \u0645\u06cc \u06a9\u0646\u062f \u0631\u0648\u06cc \u0645\u062c\u0645\u0648\u0639\u0647 \u062a\u0633\u062a:<\/p>\n<pre><code class=\"hljs\">score = model.evaluate(X_test, y_test, verbose=<span class=\"hljs-number\">1<\/span>)\n\n<span class=\"hljs-built_in\">print<\/span>(<span class=\"hljs-string\">\"Test Score:\"<\/span>, score(<span class=\"hljs-number\">0<\/span>))\n<span class=\"hljs-built_in\">print<\/span>(<span class=\"hljs-string\">\"Test Accuracy:\"<\/span>, score(<span class=\"hljs-number\">1<\/span>))\n<\/code><\/pre>\n<pre><code class=\"hljs\">10000\/10000 (==============================) - 0s 34us\/step\nTest Score: 0.8206425309181213\nTest Accuracy: 0.6669\n<\/code><\/pre>\n<p>\u0645\u0627 \u0645\u06cc\u062a\u0648\u0627\u0646\u06cc\u0645 print \u0645\u0642\u0627\u062f\u06cc\u0631 \u0627\u0632 \u062f\u0633\u062a \u062f\u0627\u062f\u0646 \u0648 \u062f\u0642\u062a \u0628\u0631\u0627\u06cc \u0645\u062c\u0645\u0648\u0639\u0647 \u0647\u0627\u06cc \u0622\u0645\u0648\u0632\u0634\u06cc \u0648 \u0622\u0632\u0645\u0627\u06cc\u0634\u06cc \u0627\u0632 \u0637\u0631\u06cc\u0642 \u0627\u0633\u06a9\u0631\u06cc\u067e\u062a \u0632\u06cc\u0631:<\/p>\n<pre><code class=\"hljs\"><span class=\"hljs-keyword\">import<\/span> matplotlib.pyplot <span class=\"hljs-keyword\">as<\/span> plt\n\nplt.plot(history.history(<span class=\"hljs-string\">'acc'<\/span>))\nplt.plot(history.history(<span class=\"hljs-string\">'val_acc'<\/span>))\n\nplt.title(<span class=\"hljs-string\">'model accuracy'<\/span>)\nplt.ylabel(<span class=\"hljs-string\">'accuracy'<\/span>)\nplt.xlabel(<span class=\"hljs-string\">'epoch'<\/span>)\nplt.legend((<span class=\"hljs-string\">'train'<\/span>,<span class=\"hljs-string\">'test'<\/span>), loc=<span class=\"hljs-string\">'upper left'<\/span>)\nplt.show()\n\nplt.plot(history.history(<span class=\"hljs-string\">'loss'<\/span>))\nplt.plot(history.history(<span class=\"hljs-string\">'val_loss'<\/span>))\n\nplt.title(<span class=\"hljs-string\">'model loss'<\/span>)\nplt.ylabel(<span class=\"hljs-string\">'loss'<\/span>)\nplt.xlabel(<span class=\"hljs-string\">'epoch'<\/span>)\nplt.legend((<span class=\"hljs-string\">'train'<\/span>,<span class=\"hljs-string\">'test'<\/span>), loc=<span class=\"hljs-string\">'upper left'<\/span>)\nplt.show()\n<\/code><\/pre>\n<p><img decoding=\"async\" class=\"img-responsive\" src=\"https:\/\/rasanegar.com\/blog\/wp-content\/uploads\/2024\/01\/python-for-nlp-creating-text-classification-model-with-multiple-inputs-in-keras-9.png\" alt=\"\u0633\u0631\" title=\"\"><\/p>\n<p>\u0627\u0632 \u062e\u0631\u0648\u062c\u06cc\u060c \u0645\u06cc \u062a\u0648\u0627\u0646\u06cc\u062f \u0628\u0628\u06cc\u0646\u06cc\u062f \u06a9\u0647 \u0645\u0642\u0627\u062f\u06cc\u0631 \u062f\u0642\u062a \u0646\u0633\u0628\u062a\u0627\u064b \u06a9\u0645\u062a\u0631 \u0627\u0633\u062a.  \u0627\u0632 \u0627\u06cc\u0646 \u0631\u0648\u060c \u0645\u06cc \u062a\u0648\u0627\u0646 \u06af\u0641\u062a \u06a9\u0647 \u0645\u062f\u0644 \u0645\u0627 \u06a9\u0645\u062a\u0631 \u0628\u0631\u0627\u0632\u0646\u062f\u0647 \u0627\u0633\u062a.  \u062f\u0642\u062a \u0631\u0627 \u0645\u06cc \u062a\u0648\u0627\u0646 \u0628\u0627 \u0627\u0641\u0632\u0627\u06cc\u0634 \u062a\u0639\u062f\u0627\u062f \u0644\u0627\u06cc\u0647 \u0647\u0627\u06cc \u0645\u062a\u0631\u0627\u06a9\u0645 \u06cc\u0627 \u0628\u0627 \u0627\u0641\u0632\u0627\u06cc\u0634 \u062a\u0639\u062f\u0627\u062f \u062f\u0648\u0631\u0647 \u0647\u0627 \u0627\u0641\u0632\u0627\u06cc\u0634 \u062f\u0627\u062f\u060c \u0628\u0627 \u0627\u06cc\u0646 \u062d\u0627\u0644 \u0645\u0646 \u0622\u0646 \u0631\u0627 \u0628\u0647 \u0634\u0645\u0627 \u0648\u0627\u06af\u0630\u0627\u0631 \u0645\u06cc \u06a9\u0646\u0645.<\/p>\n<p>\u0628\u06cc\u0627 \u062d\u0631\u06a9\u062a \u06a9\u0646\u06cc\u0645 \u0631\u0648\u06cc \u0628\u0647 \u0622\u062e\u0631\u06cc\u0646 \u0648 \u0645\u0647\u0645\u062a\u0631\u06cc\u0646 \u0628\u062e\u0634 \u0627\u06cc\u0646 \u0645\u0642\u0627\u0644\u0647 \u06a9\u0647 \u062f\u0631 \u0622\u0646 \u0627\u0632 \u0686\u0646\u062f\u06cc\u0646 \u0648\u0631\u0648\u062f\u06cc \u0627\u0632 \u0627\u0646\u0648\u0627\u0639 \u0645\u062e\u062a\u0644\u0641 \u0628\u0631\u0627\u06cc \u0622\u0645\u0648\u0632\u0634 \u0645\u062f\u0644 \u062e\u0648\u062f \u0627\u0633\u062a\u0641\u0627\u062f\u0647 \u062e\u0648\u0627\u0647\u06cc\u0645 \u06a9\u0631\u062f.<\/p>\n<h2 id=\"creatingamodelwithmultipleinputs\"><span class=\"ez-toc-section\" id=\"%d8%a7%db%8c%d8%ac%d8%a7%d8%af_%db%8c%da%a9_%d9%85%d8%af%d9%84_%d8%a8%d8%a7_%d9%88%d8%b1%d9%88%d8%af%db%8c_%d9%87%d8%a7%db%8c_%d9%85%d8%aa%d8%b9%d8%af%d8%af\"><\/span>\u0627\u06cc\u062c\u0627\u062f \u06cc\u06a9 \u0645\u062f\u0644 \u0628\u0627 \u0648\u0631\u0648\u062f\u06cc \u0647\u0627\u06cc \u0645\u062a\u0639\u062f\u062f<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<p>\u062f\u0631 \u0628\u062e\u0634\u200c\u0647\u0627\u06cc \u0642\u0628\u0644\u06cc\u060c \u0631\u0648\u0634 \u0622\u0645\u0648\u0632\u0634 \u0645\u062f\u0644\u200c\u0647\u0627\u06cc \u06cc\u0627\u062f\u06af\u06cc\u0631\u06cc \u0639\u0645\u06cc\u0642 \u0628\u0627 \u0627\u0633\u062a\u0641\u0627\u062f\u0647 \u0627\u0632 \u062f\u0627\u062f\u0647\u200c\u0647\u0627\u06cc \u0645\u062a\u0646\u06cc \u06cc\u0627 \u0645\u062a\u0627 \u0631\u0627 \u062f\u06cc\u062f\u06cc\u0645.  \u0627\u06af\u0631 \u0628\u062e\u0648\u0627\u0647\u06cc\u0645 \u0627\u0637\u0644\u0627\u0639\u0627\u062a \u0645\u062a\u0646\u06cc \u0631\u0627 \u0628\u0627 \u0647\u0645 \u062a\u0631\u06a9\u06cc\u0628 \u06a9\u0646\u06cc\u0645 \u0686\u0647\u061f <em>\u0628\u0627<\/em> \u0627\u0637\u0644\u0627\u0639\u0627\u062a \u0645\u062a\u0627 \u0648 \u0627\u0633\u062a\u0641\u0627\u062f\u0647 \u0627\u0632 \u0622\u0646 \u0628\u0647 \u0639\u0646\u0648\u0627\u0646 \u0648\u0631\u0648\u062f\u06cc \u0645\u062f\u0644 \u062e\u0648\u062f\u061f  \u0645\u0627 \u0645\u06cc \u062a\u0648\u0627\u0646\u06cc\u0645 \u0627\u06cc\u0646 \u06a9\u0627\u0631 \u0631\u0627 \u0628\u0627 \u0627\u0633\u062a\u0641\u0627\u062f\u0647 \u0627\u0632 API \u0639\u0645\u0644\u06a9\u0631\u062f\u06cc Keras \u0627\u0646\u062c\u0627\u0645 \u062f\u0647\u06cc\u0645.  \u062f\u0631 \u0627\u06cc\u0646 \u0642\u0633\u0645\u062a \u062f\u0648 \u0645\u062f\u0644 \u0641\u0631\u0639\u06cc \u0627\u06cc\u062c\u0627\u062f \u0645\u06cc \u06a9\u0646\u06cc\u0645.<\/p>\n<p>\u0627\u0648\u0644\u06cc\u0646 \u0645\u062f\u0644 \u0641\u0631\u0639\u06cc \u0648\u0631\u0648\u062f\u06cc \u0645\u062a\u0646\u06cc \u0631\u0627 \u062f\u0631 \u0642\u0627\u0644\u0628 \u0628\u0631\u0631\u0633\u06cc \u0645\u062a\u0646\u06cc \u0645\u06cc \u067e\u0630\u06cc\u0631\u062f.  \u0627\u06cc\u0646 \u0645\u062f\u0644 \u0641\u0631\u0639\u06cc \u0627\u0632 \u06cc\u06a9 \u0644\u0627\u06cc\u0647 \u0634\u06a9\u0644 \u0648\u0631\u0648\u062f\u06cc\u060c \u06cc\u06a9 \u0644\u0627\u06cc\u0647 \u062c\u0627\u0633\u0627\u0632\u06cc \u0648 \u06cc\u06a9 \u0644\u0627\u06cc\u0647 LSTM \u0627\u0632 128 \u0646\u0648\u0631\u0648\u0646 \u062a\u0634\u06a9\u06cc\u0644 \u0634\u062f\u0647 \u0627\u0633\u062a.  \u0645\u062f\u0644 \u0641\u0631\u0639\u06cc \u062f\u0648\u0645 \u0648\u0631\u0648\u062f\u06cc \u0631\u0627 \u0628\u0647 \u0634\u06a9\u0644 \u0645\u062a\u0627 \u0627\u0637\u0644\u0627\u0639\u0627\u062a \u0627\u0632 \u0637\u0631\u0641 \u0645\u06cc \u067e\u0630\u06cc\u0631\u062f <code>useful<\/code>\u060c <code>funny<\/code>\u060c \u0648 <code>cool<\/code> \u0633\u062a\u0648\u0646 \u0647\u0627.  \u0645\u062f\u0644 \u0641\u0631\u0639\u06cc \u062f\u0648\u0645 \u0646\u06cc\u0632 \u0627\u0632 \u0633\u0647 \u0644\u0627\u06cc\u0647 \u062a\u0634\u06a9\u06cc\u0644 \u0634\u062f\u0647 \u0627\u0633\u062a.  \u06cc\u06a9 \u0644\u0627\u06cc\u0647 \u0648\u0631\u0648\u062f\u06cc \u0648 \u062f\u0648 \u0644\u0627\u06cc\u0647 \u0645\u062a\u0631\u0627\u06a9\u0645.<\/p>\n<p>\u062e\u0631\u0648\u062c\u06cc \u0644\u0627\u06cc\u0647 LSTM \u0632\u06cc\u0631\u0645\u062f\u0644 \u0627\u0648\u0644 \u0648 \u062e\u0631\u0648\u062c\u06cc \u0644\u0627\u06cc\u0647 \u0645\u062a\u0631\u0627\u06a9\u0645 \u062f\u0648\u0645 \u0627\u0632 \u0632\u06cc\u0631\u0645\u062f\u0644 \u062f\u0648\u0645 \u0628\u0647 \u06cc\u06a9\u062f\u06cc\u06af\u0631 \u0645\u062a\u0635\u0644 \u0645\u06cc \u0634\u0648\u0646\u062f \u0648 \u0628\u0647 \u0639\u0646\u0648\u0627\u0646 \u0648\u0631\u0648\u062f\u06cc \u067e\u06cc\u0648\u0633\u062a\u0647 \u0628\u0647 \u0644\u0627\u06cc\u0647 \u0645\u062a\u0631\u0627\u06a9\u0645 \u062f\u06cc\u06af\u0631\u06cc \u0628\u0627 10 \u0646\u0648\u0631\u0648\u0646 \u0627\u0633\u062a\u0641\u0627\u062f\u0647 \u0645\u06cc \u0634\u0648\u0646\u062f.  \u062f\u0631 \u0646\u0647\u0627\u06cc\u062a\u060c \u0644\u0627\u06cc\u0647 \u0645\u062a\u0631\u0627\u06a9\u0645 \u062e\u0631\u0648\u062c\u06cc \u062f\u0627\u0631\u0627\u06cc \u0633\u0647 \u0646\u0648\u0631\u0648\u0646 \u0645\u0631\u0628\u0648\u0637 \u0628\u0647 \u0647\u0631 \u0646\u0648\u0639 \u0628\u0631\u0631\u0633\u06cc \u062e\u0648\u0627\u0647\u062f \u0628\u0648\u062f.<\/p>\n<p>\u0628\u06cc\u0627\u06cc\u06cc\u062f \u0628\u0628\u06cc\u0646\u06cc\u0645 \u0686\u06af\u0648\u0646\u0647 \u0645\u06cc \u062a\u0648\u0627\u0646\u06cc\u0645 \u0686\u0646\u06cc\u0646 \u0645\u062f\u0644 \u067e\u06cc\u0648\u0633\u062a\u0647 \u0627\u06cc \u0627\u06cc\u062c\u0627\u062f \u06a9\u0646\u06cc\u0645.<\/p>\n<p>\u0627\u0628\u062a\u062f\u0627 \u0628\u0627\u06cc\u062f \u062f\u0648 \u0646\u0648\u0639 \u0648\u0631\u0648\u062f\u06cc \u0645\u062e\u062a\u0644\u0641 \u0627\u06cc\u062c\u0627\u062f \u06a9\u0646\u06cc\u0645.  \u0628\u0631\u0627\u06cc \u0627\u0646\u062c\u0627\u0645 \u0627\u06cc\u0646 \u06a9\u0627\u0631\u060c \u0645\u0627\u0646\u0646\u062f \u0634\u06a9\u0644 \u0632\u06cc\u0631\u060c \u062f\u0627\u062f\u0647 \u0647\u0627\u06cc \u062e\u0648\u062f \u0631\u0627 \u0628\u0647 \u06cc\u06a9 \u0645\u062c\u0645\u0648\u0639\u0647 \u0648\u06cc\u0698\u06af\u06cc \u0648 \u0645\u062c\u0645\u0648\u0639\u0647 \u0628\u0631\u0686\u0633\u0628 \u062a\u0642\u0633\u06cc\u0645 \u0645\u06cc \u06a9\u0646\u06cc\u0645:<\/p>\n<pre><code class=\"hljs\">X = yelp_reviews.drop(<span class=\"hljs-string\">'reviews_score'<\/span>, axis=<span class=\"hljs-number\">1<\/span>)\n\ny = yelp_reviews(<span class=\"hljs-string\">'reviews_score'<\/span>)\n<\/code><\/pre>\n<p>\u0627\u06cc\u0646 <code>X<\/code> \u0645\u062a\u063a\u06cc\u0631 \u0634\u0627\u0645\u0644 \u0645\u062c\u0645\u0648\u0639\u0647 \u0648\u06cc\u0698\u06af\u06cc \u0627\u0633\u062a\u060c \u062f\u0631 \u062d\u0627\u0644\u06cc \u06a9\u0647 <code>y<\/code> \u0645\u062a\u063a\u06cc\u0631 \u0634\u0627\u0645\u0644 \u0645\u062c\u0645\u0648\u0639\u0647 \u0628\u0631\u0686\u0633\u0628 \u0627\u0633\u062a.  \u0645\u0627 \u0628\u0627\u06cc\u062f \u0628\u0631\u0686\u0633\u0628 \u0647\u0627\u06cc \u062e\u0648\u062f \u0631\u0627 \u0628\u0647 \u0628\u0631\u062f\u0627\u0631\u0647\u0627\u06cc \u0631\u0645\u0632\u06af\u0630\u0627\u0631\u06cc \u0634\u062f\u0647 \u06cc\u06a9 \u062f\u0627\u063a \u062a\u0628\u062f\u06cc\u0644 \u06a9\u0646\u06cc\u0645.  \u0645\u0627 \u0645\u06cc \u062a\u0648\u0627\u0646\u06cc\u0645 \u0627\u06cc\u0646 \u06a9\u0627\u0631 \u0631\u0627 \u0628\u0627 \u0627\u0633\u062a\u0641\u0627\u062f\u0647 \u0627\u0632 \u0631\u0645\u0632\u06af\u0630\u0627\u0631 \u0628\u0631\u0686\u0633\u0628 \u0648 \u06a9\u062f \u0627\u0646\u062c\u0627\u0645 \u062f\u0647\u06cc\u0645 <code>to_categorical<\/code> \u0639\u0645\u0644\u06a9\u0631\u062f \u0627\u0632 <code>keras.utils<\/code> \u0645\u062f\u0648\u0644.  \u0645\u0627 \u0647\u0645\u0686\u0646\u06cc\u0646 \u062f\u0627\u062f\u0647 \u0647\u0627\u06cc \u062e\u0648\u062f \u0631\u0627 \u0628\u0647 \u0645\u062c\u0645\u0648\u0639\u0647 \u0647\u0627\u06cc \u0622\u0645\u0648\u0632\u0634\u06cc \u0648 \u0648\u06cc\u0698\u06af\u06cc \u0647\u0627 \u062a\u0642\u0633\u06cc\u0645 \u0645\u06cc \u06a9\u0646\u06cc\u0645.<\/p>\n<pre><code class=\"hljs\"><span class=\"hljs-keyword\">from<\/span> sklearn <span class=\"hljs-keyword\">import<\/span> preprocessing\n\n\nlabel_encoder = preprocessing.LabelEncoder()\n\n\ny = label_encoder.fit_transform(y)\n\nX_train, X_test, y_train, y_test = train_test_split(X, y, test_size=<span class=\"hljs-number\">0.20<\/span>, random_state=<span class=\"hljs-number\">42<\/span>)\n\n<span class=\"hljs-keyword\">from<\/span> keras.utils <span class=\"hljs-keyword\">import<\/span> to_categorical\ny_train = to_categorical(y_train)\ny_test = to_categorical(y_test)\n<\/code><\/pre>\n<p>\u0627\u06a9\u0646\u0648\u0646 \u0645\u062c\u0645\u0648\u0639\u0647 \u0628\u0631\u0686\u0633\u0628 \u0645\u0627 \u0628\u0647 \u0634\u06a9\u0644 \u0645\u0648\u0631\u062f \u0646\u06cc\u0627\u0632 \u0627\u0633\u062a.  \u0627\u0632 \u0622\u0646\u062c\u0627\u06cc\u06cc \u06a9\u0647 \u062a\u0646\u0647\u0627 \u06cc\u06a9 \u062e\u0631\u0648\u062c\u06cc \u0648\u062c\u0648\u062f \u062e\u0648\u0627\u0647\u062f \u062f\u0627\u0634\u062a\u060c \u0628\u0646\u0627\u0628\u0631\u0627\u06cc\u0646 \u0646\u06cc\u0627\u0632\u06cc \u0628\u0647 \u0622\u0646 \u0646\u062f\u0627\u0631\u06cc\u0645 process \u0645\u062c\u0645\u0648\u0639\u0647 \u0628\u0631\u0686\u0633\u0628 \u0645\u0627  \u0628\u0627 \u0627\u06cc\u0646 \u062d\u0627\u0644\u060c \u0686\u0646\u062f\u06cc\u0646 \u0648\u0631\u0648\u062f\u06cc \u0628\u0631\u0627\u06cc \u0645\u062f\u0644 \u0648\u062c\u0648\u062f \u062e\u0648\u0627\u0647\u062f \u062f\u0627\u0634\u062a.  \u0628\u0646\u0627\u0628\u0631\u0627\u06cc\u0646\u060c \u0645\u0627 \u0628\u0627\u06cc\u062f \u0645\u062c\u0645\u0648\u0639\u0647 \u0648\u06cc\u0698\u06af\u06cc \u0647\u0627\u06cc \u062e\u0648\u062f \u0631\u0627 \u0627\u0632 \u0642\u0628\u0644 \u067e\u0631\u062f\u0627\u0632\u0634 \u06a9\u0646\u06cc\u0645.<\/p>\n<p>\u0627\u0628\u062a\u062f\u0627 \u0628\u06cc\u0627\u06cc\u06cc\u062f \u0627\u06cc\u062c\u0627\u062f \u06a9\u0646\u06cc\u0645 <code>preproces_text<\/code> \u062a\u0627\u0628\u0639\u06cc \u06a9\u0647 \u0628\u0631\u0627\u06cc \u067e\u06cc\u0634 \u067e\u0631\u062f\u0627\u0632\u0634 \u0645\u062c\u0645\u0648\u0639\u0647 \u062f\u0627\u062f\u0647 \u0645\u0627 \u0627\u0633\u062a\u0641\u0627\u062f\u0647 \u0645\u06cc \u0634\u0648\u062f:<\/p>\n<pre><code class=\"hljs\"><span class=\"hljs-function\"><span class=\"hljs-keyword\">def<\/span> <span class=\"hljs-title\">preprocess_text<\/span>(<span class=\"hljs-params\">sen<\/span>):<\/span>\n\n    \n    sentence = re.sub(<span class=\"hljs-string\">'(^a-zA-Z)'<\/span>, <span class=\"hljs-string\">' '<\/span>, sen)\n\n    \n    sentence = re.sub(<span class=\"hljs-string\">r\"\\s+(a-zA-Z)\\s+\"<\/span>, <span class=\"hljs-string\">' '<\/span>, sentence)\n\n    \n    sentence = re.sub(<span class=\"hljs-string\">r'\\s+'<\/span>, <span class=\"hljs-string\">' '<\/span>, sentence)\n\n    <span class=\"hljs-keyword\">return<\/span> sentence\n<\/code><\/pre>\n<p>\u0628\u0647 \u0639\u0646\u0648\u0627\u0646 \u0627\u0648\u0644\u06cc\u0646 \u0642\u062f\u0645\u060c \u0648\u0631\u0648\u062f\u06cc \u0645\u062a\u0646\u06cc \u0631\u0627 \u0628\u0631\u0627\u06cc \u0645\u062c\u0645\u0648\u0639\u0647 \u0622\u0645\u0648\u0632\u0634\u06cc \u0648 \u062a\u0633\u062a \u0627\u06cc\u062c\u0627\u062f \u0645\u06cc \u06a9\u0646\u06cc\u0645.  \u0628\u0647 \u0627\u0633\u06a9\u0631\u06cc\u067e\u062a \u0632\u06cc\u0631 \u0646\u06af\u0627\u0647 \u06a9\u0646\u06cc\u062f:<\/p>\n<pre><code class=\"hljs\">X1_train = ()\nsentences = <span class=\"hljs-built_in\">list<\/span>(X_train(<span class=\"hljs-string\">\"text\"<\/span>))\n<span class=\"hljs-keyword\">for<\/span> sen <span class=\"hljs-keyword\">in<\/span> sentences:\n    X1_train.append(preprocess_text(sen))\n<\/code><\/pre>\n<p>\u0627\u06a9\u0646\u0648\u0646 <code>X1_train<\/code> \u062d\u0627\u0648\u06cc \u0648\u0631\u0648\u062f\u06cc \u0645\u062a\u0646\u06cc \u0645\u062c\u0645\u0648\u0639\u0647 \u0622\u0645\u0648\u0632\u0634\u06cc \u0627\u0633\u062a.  \u0628\u0647 \u0637\u0648\u0631 \u0645\u0634\u0627\u0628\u0647\u060c \u0627\u0633\u06a9\u0631\u06cc\u067e\u062a \u0632\u06cc\u0631 \u062f\u0627\u062f\u0647 \u0647\u0627\u06cc \u0648\u0631\u0648\u062f\u06cc \u0645\u062a\u0646\u06cc \u0631\u0627 \u0628\u0631\u0627\u06cc \u0645\u062c\u0645\u0648\u0639\u0647 \u0622\u0632\u0645\u0627\u06cc\u0634\u06cc \u067e\u06cc\u0634 \u067e\u0631\u062f\u0627\u0632\u0634 \u0645\u06cc \u06a9\u0646\u062f:<\/p>\n<pre><code class=\"hljs\">X1_test = ()\nsentences = <span class=\"hljs-built_in\">list<\/span>(X_test(<span class=\"hljs-string\">\"text\"<\/span>))\n<span class=\"hljs-keyword\">for<\/span> sen <span class=\"hljs-keyword\">in<\/span> sentences:\n    X1_test.append(preprocess_text(sen))\n<\/code><\/pre>\n<p>\u0627\u06a9\u0646\u0648\u0646 \u0628\u0627\u06cc\u062f \u0648\u0631\u0648\u062f\u06cc \u0645\u062a\u0646\u06cc \u0645\u062c\u0645\u0648\u0639\u0647 \u0647\u0627\u06cc \u0622\u0645\u0648\u0632\u0634\u06cc \u0648 \u062a\u0633\u062a \u0631\u0627 \u0628\u0627 \u0627\u0633\u062a\u0641\u0627\u062f\u0647 \u0627\u0632 \u062c\u0627\u0633\u0627\u0632\u06cc \u06a9\u0644\u0645\u0647 \u0628\u0647 \u0634\u06a9\u0644 \u0639\u062f\u062f\u06cc \u062a\u0628\u062f\u06cc\u0644 \u06a9\u0646\u06cc\u0645.  \u0627\u0633\u06a9\u0631\u06cc\u067e\u062a \u0632\u06cc\u0631 \u0627\u06cc\u0646 \u06a9\u0627\u0631 \u0631\u0627 \u0627\u0646\u062c\u0627\u0645 \u0645\u06cc \u062f\u0647\u062f:<\/p>\n<pre><code class=\"hljs\">tokenizer = Tokenizer(num_words=<span class=\"hljs-number\">5000<\/span>)\ntokenizer.fit_on_texts(X1_train)\n\nX1_train = tokenizer.texts_to_sequences(X1_train)\nX1_test = tokenizer.texts_to_sequences(X1_test)\n\nvocab_size = <span class=\"hljs-built_in\">len<\/span>(tokenizer.word_index) + <span class=\"hljs-number\">1<\/span>\n\nmaxlen = <span class=\"hljs-number\">200<\/span>\n\nX1_train = pad_sequences(X1_train, padding=<span class=\"hljs-string\">'post'<\/span>, maxlen=maxlen)\nX1_test = pad_sequences(X1_test, padding=<span class=\"hljs-string\">'post'<\/span>, maxlen=maxlen)\n<\/code><\/pre>\n<p>\u0645\u0627 \u062f\u0648\u0628\u0627\u0631\u0647 \u0627\u0632 \u062c\u0627\u0633\u0627\u0632\u06cc \u06a9\u0644\u0645\u0647 GloVe \u0628\u0631\u0627\u06cc \u0627\u06cc\u062c\u0627\u062f \u0628\u0631\u062f\u0627\u0631 \u06a9\u0644\u0645\u0627\u062a \u0627\u0633\u062a\u0641\u0627\u062f\u0647 \u062e\u0648\u0627\u0647\u06cc\u0645 \u06a9\u0631\u062f:<\/p>\n<pre><code class=\"hljs\"><span class=\"hljs-keyword\">from<\/span> numpy <span class=\"hljs-keyword\">import<\/span> array\n<span class=\"hljs-keyword\">from<\/span> numpy <span class=\"hljs-keyword\">import<\/span> asarray\n<span class=\"hljs-keyword\">from<\/span> numpy <span class=\"hljs-keyword\">import<\/span> zeros\n\nembeddings_dictionary = <span class=\"hljs-built_in\">dict<\/span>()\n\nglove_file = <span class=\"hljs-built_in\">open<\/span>(<span class=\"hljs-string\">'\/content\/drive\/My Drive\/glove.6B.100d.txt'<\/span>, encoding=<span class=\"hljs-string\">\"utf8\"<\/span>)\n\n<span class=\"hljs-keyword\">for<\/span> line <span class=\"hljs-keyword\">in<\/span> glove_file:\n    records = line.split()\n    word = records(<span class=\"hljs-number\">0<\/span>)\n    vector_dimensions = asarray(records(<span class=\"hljs-number\">1<\/span>:), dtype=<span class=\"hljs-string\">'float32'<\/span>)\n    embeddings_dictionary(word) = vector_dimensions\n\nglove_file.close()\n\nembedding_matrix = zeros((vocab_size, <span class=\"hljs-number\">100<\/span>))\n<span class=\"hljs-keyword\">for<\/span> word, index <span class=\"hljs-keyword\">in<\/span> tokenizer.word_index.items():\n    embedding_vector = embeddings_dictionary.get(word)\n    <span class=\"hljs-keyword\">if<\/span> embedding_vector <span class=\"hljs-keyword\">is<\/span> <span class=\"hljs-keyword\">not<\/span> <span class=\"hljs-literal\">None<\/span>:\n        embedding_matrix(index) = embedding_vector\n<\/code><\/pre>\n<p>\u0645\u0627 \u0648\u0631\u0648\u062f\u06cc \u0645\u062a\u0646\u06cc \u062e\u0648\u062f \u0631\u0627 \u0627\u0632 \u0642\u0628\u0644 \u067e\u0631\u062f\u0627\u0632\u0634 \u06a9\u0631\u062f\u0647 \u0627\u06cc\u0645.  \u0646\u0648\u0639 \u062f\u0648\u0645 \u0648\u0631\u0648\u062f\u06cc \u0645\u062a\u0627 \u0627\u0637\u0644\u0627\u0639\u0627\u062a \u0645\u0648\u062c\u0648\u062f \u062f\u0631 <code>useful<\/code>\u060c <code>funny<\/code>\u060c \u0648 <code>cool<\/code> \u0633\u062a\u0648\u0646 \u0647\u0627.  \u0645\u0627 \u0627\u06cc\u0646 \u0633\u062a\u0648\u0646 \u0647\u0627 \u0631\u0627 \u0627\u0632 \u0645\u062c\u0645\u0648\u0639\u0647 \u0648\u06cc\u0698\u06af\u06cc \u0641\u06cc\u0644\u062a\u0631 \u0645\u06cc \u06a9\u0646\u06cc\u0645 \u062a\u0627 \u0648\u0631\u0648\u062f\u06cc \u0645\u062a\u0627 \u0628\u0631\u0627\u06cc \u0622\u0645\u0648\u0632\u0634 \u0627\u0644\u06af\u0648\u0631\u06cc\u062a\u0645 \u0647\u0627 \u0627\u06cc\u062c\u0627\u062f \u06a9\u0646\u06cc\u0645.  \u0628\u0647 \u0627\u0633\u06a9\u0631\u06cc\u067e\u062a \u0632\u06cc\u0631 \u0646\u06af\u0627\u0647 \u06a9\u0646\u06cc\u062f:<\/p>\n<pre><code class=\"hljs\">X2_train = X_train((<span class=\"hljs-string\">'useful'<\/span>, <span class=\"hljs-string\">'funny'<\/span>, <span class=\"hljs-string\">'cool'<\/span>)).values\nX2_test = X_test((<span class=\"hljs-string\">'useful'<\/span>, <span class=\"hljs-string\">'funny'<\/span>, <span class=\"hljs-string\">'cool'<\/span>)).values\n<\/code><\/pre>\n<p>\u062d\u0627\u0644 \u0628\u06cc\u0627\u06cc\u06cc\u062f \u062f\u0648 \u0644\u0627\u06cc\u0647 \u0648\u0631\u0648\u062f\u06cc \u062e\u0648\u062f \u0631\u0627 \u0627\u06cc\u062c\u0627\u062f \u06a9\u0646\u06cc\u0645.  \u0644\u0627\u06cc\u0647 \u0648\u0631\u0648\u062f\u06cc \u0627\u0648\u0644 \u0628\u0631\u0627\u06cc \u0648\u0627\u0631\u062f \u06a9\u0631\u062f\u0646 \u0648\u0631\u0648\u062f\u06cc \u0645\u062a\u0646\u06cc \u0648 \u0644\u0627\u06cc\u0647 \u0648\u0631\u0648\u062f\u06cc \u062f\u0648\u0645 \u0628\u0631\u0627\u06cc \u0648\u0631\u0648\u062f \u0627\u0637\u0644\u0627\u0639\u0627\u062a \u0645\u062a\u0627 \u0627\u0632 \u0633\u0647 \u0633\u062a\u0648\u0646 \u0627\u0633\u062a\u0641\u0627\u062f\u0647 \u0645\u06cc \u0634\u0648\u062f.<\/p>\n<pre><code class=\"hljs\">input_1 = Input(shape=(maxlen,))\n\ninput_2 = Input(shape=(<span class=\"hljs-number\">3<\/span>,))\n<\/code><\/pre>\n<p>\u0645\u06cc \u0628\u06cc\u0646\u06cc\u062f \u06a9\u0647 \u0627\u0648\u0644\u06cc\u0646 \u0644\u0627\u06cc\u0647 \u0648\u0631\u0648\u062f\u06cc \u0627\u0633\u062a <code>input_1<\/code> \u0628\u0631\u0627\u06cc \u0648\u0631\u0648\u062f\u06cc \u0645\u062a\u0646\u06cc \u0627\u0633\u062a\u0641\u0627\u062f\u0647 \u0645\u06cc \u0634\u0648\u062f.  \u0627\u0646\u062f\u0627\u0632\u0647 \u0634\u06a9\u0644 \u0628\u0647 \u0634\u06a9\u0644 \u062c\u0645\u0644\u0647 \u0648\u0631\u0648\u062f\u06cc \u062a\u0646\u0638\u06cc\u0645 \u0634\u062f\u0647 \u0627\u0633\u062a.  \u0628\u0631\u0627\u06cc \u0644\u0627\u06cc\u0647 \u0648\u0631\u0648\u062f\u06cc \u062f\u0648\u0645\u060c \u0634\u06a9\u0644 \u0645\u0631\u0628\u0648\u0637 \u0628\u0647 \u0633\u0647 \u0633\u062a\u0648\u0646 \u0627\u0633\u062a.<\/p>\n<p>\u0628\u06cc\u0627\u06cc\u06cc\u062f \u0627\u06a9\u0646\u0648\u0646 \u0627\u0648\u0644\u06cc\u0646 \u0645\u062f\u0644 \u0641\u0631\u0639\u06cc \u0631\u0627 \u0627\u06cc\u062c\u0627\u062f \u06a9\u0646\u06cc\u0645 \u06a9\u0647 \u062f\u0627\u062f\u0647 \u0647\u0627 \u0631\u0627 \u0627\u0632 \u0627\u0648\u0644\u06cc\u0646 \u0644\u0627\u06cc\u0647 \u0648\u0631\u0648\u062f\u06cc \u0645\u06cc \u067e\u0630\u06cc\u0631\u062f:<\/p>\n<pre><code class=\"hljs\">embedding_layer = Embedding(vocab_size, <span class=\"hljs-number\">100<\/span>, weights=(embedding_matrix), trainable=<span class=\"hljs-literal\">False<\/span>)(input_1)\nLSTM_Layer_1 = LSTM(<span class=\"hljs-number\">128<\/span>)(embedding_layer)\n<\/code><\/pre>\n<p>\u0628\u0647 \u0637\u0648\u0631 \u0645\u0634\u0627\u0628\u0647\u060c \u0627\u0633\u06a9\u0631\u06cc\u067e\u062a \u0632\u06cc\u0631 \u06cc\u06a9 \u0645\u062f\u0644 \u0641\u0631\u0639\u06cc \u062f\u0648\u0645 \u0627\u06cc\u062c\u0627\u062f \u0645\u06cc \u06a9\u0646\u062f \u06a9\u0647 \u0648\u0631\u0648\u062f\u06cc \u0644\u0627\u06cc\u0647 \u0648\u0631\u0648\u062f\u06cc \u062f\u0648\u0645 \u0631\u0627 \u0645\u06cc \u067e\u0630\u06cc\u0631\u062f:<\/p>\n<pre><code class=\"hljs\">dense_layer_1 = Dense(<span class=\"hljs-number\">10<\/span>, activation=<span class=\"hljs-string\">'relu'<\/span>)(input_2)\ndense_layer_2 = Dense(<span class=\"hljs-number\">10<\/span>, activation=<span class=\"hljs-string\">'relu'<\/span>)(dense_layer_1)\n<\/code><\/pre>\n<p>\u0627\u06a9\u0646\u0648\u0646 \u062f\u0648 \u0645\u062f\u0644 \u0641\u0631\u0639\u06cc \u062f\u0627\u0631\u06cc\u0645.  \u06a9\u0627\u0631\u06cc \u06a9\u0647 \u0645\u06cc \u062e\u0648\u0627\u0647\u06cc\u0645 \u0627\u0646\u062c\u0627\u0645 \u062f\u0647\u06cc\u0645 \u0627\u06cc\u0646 \u0627\u0633\u062a \u06a9\u0647 \u062e\u0631\u0648\u062c\u06cc \u0632\u06cc\u0631\u0645\u062f\u0644 \u0627\u0648\u0644 \u0631\u0627 \u0628\u0627 \u062e\u0631\u0648\u062c\u06cc \u0632\u06cc\u0631\u0645\u062f\u0644 \u062f\u0648\u0645 \u0628\u0647 \u0647\u0645 \u067e\u06cc\u0648\u0646\u062f \u062f\u0647\u06cc\u0645.  \u062e\u0631\u0648\u062c\u06cc \u0627\u0632 \u0645\u062f\u0644 \u0641\u0631\u0639\u06cc \u0627\u0648\u0644\u060c \u062e\u0631\u0648\u062c\u06cc \u0627\u0632 \u0645\u062f\u0644 \u0641\u0631\u0639\u06cc \u0627\u0633\u062a <code>LSTM_Layer_1<\/code> \u0648 \u0628\u0647 \u0637\u0648\u0631 \u0645\u0634\u0627\u0628\u0647\u060c \u062e\u0631\u0648\u062c\u06cc \u0627\u0632 \u0632\u06cc\u0631\u0645\u062f\u0644 \u062f\u0648\u0645\u060c \u062e\u0631\u0648\u062c\u06cc \u0627\u0632 \u0645\u062f\u0644 \u0641\u0631\u0639\u06cc \u0627\u0633\u062a <code>dense_layer_2<\/code>.  \u0645\u0627 \u0645\u06cc \u062a\u0648\u0627\u0646\u06cc\u0645 \u0627\u0633\u062a\u0641\u0627\u062f\u0647 \u06a9\u0646\u06cc\u0645 <code>Concatenate<\/code> \u06a9\u0644\u0627\u0633 \u0627\u0632 <code>keras.layers.merge<\/code> \u0645\u0627\u0698\u0648\u0644 \u0628\u0631\u0627\u06cc \u0628\u0647 \u0647\u0645 \u067e\u06cc\u0648\u0633\u062a\u0646 \u062f\u0648 \u0648\u0631\u0648\u062f\u06cc<\/p>\n<p>\u0627\u0633\u06a9\u0631\u06cc\u067e\u062a \u0632\u06cc\u0631 \u0645\u062f\u0644 \u0646\u0647\u0627\u06cc\u06cc \u0645\u0627 \u0631\u0627 \u0627\u06cc\u062c\u0627\u062f \u0645\u06cc \u06a9\u0646\u062f:<\/p>\n<pre><code class=\"hljs\">concat_layer = Concatenate()((LSTM_Layer_1, dense_layer_2))\ndense_layer_3 = Dense(<span class=\"hljs-number\">10<\/span>, activation=<span class=\"hljs-string\">'relu'<\/span>)(concat_layer)\noutput = Dense(<span class=\"hljs-number\">3<\/span>, activation=<span class=\"hljs-string\">'softmax'<\/span>)(dense_layer_3)\nmodel = Model(inputs=(input_1, input_2), outputs=output)\n<\/code><\/pre>\n<p>\u0645\u06cc \u0628\u06cc\u0646\u06cc\u062f \u06a9\u0647 \u0627\u06a9\u0646\u0648\u0646 \u0645\u062f\u0644 \u0645\u0627 \u0644\u06cc\u0633\u062a\u06cc \u0627\u0632 \u0648\u0631\u0648\u062f\u06cc \u0647\u0627 \u0628\u0627 \u062f\u0648 \u0622\u06cc\u062a\u0645 \u062f\u0627\u0631\u062f.  \u0627\u0633\u06a9\u0631\u06cc\u067e\u062a \u0632\u06cc\u0631 \u0645\u062f\u0644 \u0631\u0627 \u06a9\u0627\u0645\u067e\u0627\u06cc\u0644 \u0645\u06cc \u06a9\u0646\u062f \u0648 \u062e\u0644\u0627\u0635\u0647 \u0622\u0646 \u0631\u0627 \u0686\u0627\u067e \u0645\u06cc \u06a9\u0646\u062f:<\/p>\n<pre><code class=\"hljs\">model.<span class=\"hljs-built_in\">compile<\/span>(loss=<span class=\"hljs-string\">'categorical_crossentropy'<\/span>, optimizer=<span class=\"hljs-string\">'adam'<\/span>, metrics=(<span class=\"hljs-string\">'acc'<\/span>))\n<span class=\"hljs-built_in\">print<\/span>(model.summary())\n<\/code><\/pre>\n<p>\u062e\u0644\u0627\u0635\u0647 \u0645\u062f\u0644 \u0628\u0647 \u0634\u0631\u062d \u0632\u06cc\u0631 \u0627\u0633\u062a:<\/p>\n<pre><code class=\"hljs\">Layer (type)                    Output Shape         Param #     Connected to\n==================================================================================================\ninput_1 (InputLayer)            (None, 200)          0\n__________________________________________________________________________________________________\ninput_2 (InputLayer)            (None, 3)            0\n__________________________________________________________________________________________________\nembedding_1 (Embedding)         (None, 200, 100)     5572900     input_1(0)(0)\n__________________________________________________________________________________________________\ndense_1 (Dense)                 (None, 10)           40          input_2(0)(0)\n__________________________________________________________________________________________________\nlstm_1 (LSTM)                   (None, 128)          117248      embedding_1(0)(0)\n__________________________________________________________________________________________________\ndense_2 (Dense)                 (None, 10)           110         dense_1(0)(0)\n__________________________________________________________________________________________________\nconcatenate_1 (Concatenate)     (None, 138)          0           lstm_1(0)(0)\n                                                                 dense_2(0)(0)\n__________________________________________________________________________________________________\ndense_3 (Dense)                 (None, 10)           1390        concatenate_1(0)(0)\n__________________________________________________________________________________________________\ndense_4 (Dense)                 (None, 3)            33          dense_3(0)(0)\n==================================================================================================\nTotal params: 5,691,721\nTrainable params: 118,821\nNon-trainable params: 5,572,900\n<\/code><\/pre>\n<p>\u062f\u0631 \u0646\u0647\u0627\u06cc\u062a\u060c \u0645\u06cc \u062a\u0648\u0627\u0646\u06cc\u0645 \u0645\u062f\u0644 \u06a9\u0627\u0645\u0644 \u0634\u0628\u06a9\u0647 \u0631\u0627 \u0628\u0627 \u0627\u0633\u062a\u0641\u0627\u062f\u0647 \u0627\u0632 \u0627\u0633\u06a9\u0631\u06cc\u067e\u062a \u0632\u06cc\u0631 \u0631\u0633\u0645 \u06a9\u0646\u06cc\u0645:<\/p>\n<pre><code class=\"hljs\"><span class=\"hljs-keyword\">from<\/span> keras.utils <span class=\"hljs-keyword\">import<\/span> plot_model\nplot_model(model, to_file=<span class=\"hljs-string\">'model_plot3.png'<\/span>, show_shapes=<span class=\"hljs-literal\">True<\/span>, show_layer_names=<span class=\"hljs-literal\">True<\/span>)\n<\/code><\/pre>\n<p>\u0627\u06af\u0631 \u062f\u0631 \u0631\u0627 \u0628\u0627\u0632 \u06a9\u0646\u06cc\u062f <code>model_plot3.png<\/code> \u0641\u0627\u06cc\u0644\u060c \u0628\u0627\u06cc\u062f \u0646\u0645\u0648\u062f\u0627\u0631 \u0634\u0628\u06a9\u0647 \u0632\u06cc\u0631 \u0631\u0627 \u0628\u0628\u06cc\u0646\u06cc\u062f:<\/p>\n<p><img decoding=\"async\" class=\"img-responsive\" src=\"https:\/\/rasanegar.com\/blog\/wp-content\/uploads\/2024\/01\/python-for-nlp-creating-text-classification-model-with-multiple-inputs-in-keras-10.png\" alt=\"\u0633\u0631\" title=\"\"><\/p>\n<p>\u0634\u06a9\u0644 \u0628\u0627\u0644\u0627 \u0628\u0647 \u0648\u0636\u0648\u062d \u062a\u0648\u0636\u06cc\u062d \u0645\u06cc \u062f\u0647\u062f \u06a9\u0647 \u0686\u06af\u0648\u0646\u0647 \u0686\u0646\u062f\u06cc\u0646 \u0648\u0631\u0648\u062f\u06cc \u0631\u0627 \u062f\u0631 \u06cc\u06a9 \u0648\u0631\u0648\u062f\u06cc \u0628\u0631\u0627\u06cc \u0627\u06cc\u062c\u0627\u062f \u0645\u062f\u0644 \u062e\u0648\u062f \u0627\u0644\u062d\u0627\u0642 \u06a9\u0631\u062f\u0647 \u0627\u06cc\u0645.<\/p>\n<p>\u062d\u0627\u0644\u0627 \u0628\u06cc\u0627\u06cc\u06cc\u062f \u0645\u062f\u0644 \u062e\u0648\u062f \u0631\u0627 \u0622\u0645\u0648\u0632\u0634 \u062f\u0647\u06cc\u0645 \u0648 \u0646\u062a\u0627\u06cc\u062c \u0631\u0627 \u0628\u0628\u06cc\u0646\u06cc\u0645:<\/p>\n<pre><code class=\"hljs\">history = model.fit(x=(X1_train, X2_train), y=y_train, batch_size=<span class=\"hljs-number\">128<\/span>, epochs=<span class=\"hljs-number\">10<\/span>, verbose=<span class=\"hljs-number\">1<\/span>, validation_split=<span class=\"hljs-number\">0.2<\/span>)\n<\/code><\/pre>\n<p>\u062f\u0631 \u0627\u06cc\u0646\u062c\u0627 \u0646\u062a\u06cc\u062c\u0647 10 \u062f\u0648\u0631\u0647 \u0627\u0633\u062a:<\/p>\n<pre><code class=\"hljs\">Train \u0631\u0648\u06cc 32000 samples, validate \u0631\u0648\u06cc 8000 samples\nEpoch 1\/10\n32000\/32000 (==============================) - 155s 5ms\/step - loss: 0.9006 - acc: 0.6509 - val_loss: 0.8233 - val_acc: 0.6704\nEpoch 2\/10\n32000\/32000 (==============================) - 154s 5ms\/step - loss: 0.8212 - acc: 0.6670 - val_loss: 0.8141 - val_acc: 0.6745\nEpoch 3\/10\n32000\/32000 (==============================) - 154s 5ms\/step - loss: 0.8151 - acc: 0.6691 - val_loss: 0.8086 - val_acc: 0.6740\nEpoch 4\/10\n32000\/32000 (==============================) - 155s 5ms\/step - loss: 0.8121 - acc: 0.6701 - val_loss: 0.8039 - val_acc: 0.6776\nEpoch 5\/10\n32000\/32000 (==============================) - 154s 5ms\/step - loss: 0.8027 - acc: 0.6740 - val_loss: 0.7467 - val_acc: 0.6854\nEpoch 6\/10\n32000\/32000 (==============================) - 155s 5ms\/step - loss: 0.6791 - acc: 0.7158 - val_loss: 0.5764 - val_acc: 0.7560\nEpoch 7\/10\n32000\/32000 (==============================) - 154s 5ms\/step - loss: 0.5333 - acc: 0.7744 - val_loss: 0.5076 - val_acc: 0.7881\nEpoch 8\/10\n32000\/32000 (==============================) - 154s 5ms\/step - loss: 0.4857 - acc: 0.7973 - val_loss: 0.4849 - val_acc: 0.7970\nEpoch 9\/10\n32000\/32000 (==============================) - 154s 5ms\/step - loss: 0.4697 - acc: 0.8034 - val_loss: 0.4709 - val_acc: 0.8024\nEpoch 10\/10\n32000\/32000 (==============================) - 154s 5ms\/step - loss: 0.4479 - acc: 0.8123 - val_loss: 0.4592 - val_acc: 0.8079\n<\/code><\/pre>\n<p>\u0628\u0631\u0627\u06cc \u0627\u0631\u0632\u06cc\u0627\u0628\u06cc \u0645\u062f\u0644 \u062e\u0648\u062f\u060c \u0628\u0627\u06cc\u062f \u0647\u0631 \u062f\u0648 \u0648\u0631\u0648\u062f\u06cc \u062a\u0633\u062a \u0631\u0627 \u0628\u0647 \u0622\u0646 \u067e\u0627\u0633 \u06a9\u0646\u06cc\u0645 <code>evaluate<\/code> \u0639\u0645\u0644\u06a9\u0631\u062f \u0645\u0637\u0627\u0628\u0642 \u0634\u06a9\u0644 \u0632\u06cc\u0631:<\/p>\n<pre><code class=\"hljs\">score = model.evaluate(x=(X1_test, X2_test), y=y_test, verbose=<span class=\"hljs-number\">1<\/span>)\n\n<span class=\"hljs-built_in\">print<\/span>(<span class=\"hljs-string\">\"Test Score:\"<\/span>, score(<span class=\"hljs-number\">0<\/span>))\n<span class=\"hljs-built_in\">print<\/span>(<span class=\"hljs-string\">\"Test Accuracy:\"<\/span>, score(<span class=\"hljs-number\">1<\/span>))\n<\/code><\/pre>\n<p>\u062f\u0631 \u0627\u06cc\u0646\u062c\u0627 \u0646\u062a\u06cc\u062c\u0647 \u0627\u0633\u062a:<\/p>\n<pre><code class=\"hljs\">10000\/10000 (==============================) - 18s 2ms\/step\nTest Score: 0.4576087875843048\nTest Accuracy: 0.8053\n<\/code><\/pre>\n<p>\u062f\u0642\u062a \u062a\u0633\u062a \u0645\u0627 80.53\u066a \u0627\u0633\u062a \u06a9\u0647 \u06a9\u0645\u06cc \u06a9\u0645\u062a\u0631 \u0627\u0632 \u0627\u0648\u0644\u06cc\u0646 \u0645\u062f\u0644 \u0645\u0627 \u0627\u0633\u062a \u06a9\u0647 \u0641\u0642\u0637 \u0627\u0632 \u0648\u0631\u0648\u062f\u06cc \u0645\u062a\u0646\u06cc \u0627\u0633\u062a\u0641\u0627\u062f\u0647 \u0645\u06cc \u06a9\u0646\u062f.  \u0627\u06cc\u0646 \u0646\u0634\u0627\u0646 \u0645\u06cc \u062f\u0647\u062f \u06a9\u0647 \u0627\u0637\u0644\u0627\u0639\u0627\u062a \u0645\u062a\u0627 \u062f\u0631 <code>yelp_reviews<\/code> \u0628\u0631\u0627\u06cc \u067e\u06cc\u0634 \u0628\u06cc\u0646\u06cc \u0627\u062d\u0633\u0627\u0633\u0627\u062a \u062e\u06cc\u0644\u06cc \u0645\u0641\u06cc\u062f \u0646\u06cc\u0633\u062a.<\/p>\n<p>\u0628\u0647 \u0647\u0631 \u062d\u0627\u0644\u060c \u0627\u06a9\u0646\u0648\u0646 \u0645\u06cc \u062f\u0627\u0646\u06cc\u062f \u06a9\u0647 \u0686\u06af\u0648\u0646\u0647 \u0645\u062f\u0644 \u0647\u0627\u06cc \u0648\u0631\u0648\u062f\u06cc \u0686\u0646\u062f\u06af\u0627\u0646\u0647 \u0628\u0631\u0627\u06cc \u0637\u0628\u0642\u0647 \u0628\u0646\u062f\u06cc \u0645\u062a\u0646 \u062f\u0631 Keras \u0627\u06cc\u062c\u0627\u062f \u06a9\u0646\u06cc\u062f!<\/p>\n<p>\u062f\u0631 \u0646\u0647\u0627\u06cc\u062a\u060c \u0627\u062c\u0627\u0632\u0647 \u062f\u0647\u06cc\u062f \u062f\u0631 \u062d\u0627\u0644 \u062d\u0627\u0636\u0631 print \u0627\u0632 \u062f\u0633\u062a \u062f\u0627\u062f\u0646 \u0648 \u062f\u0642\u062a \u0628\u0631\u0627\u06cc \u0645\u062c\u0645\u0648\u0639\u0647 \u0647\u0627\u06cc \u0622\u0645\u0648\u0632\u0634\u06cc \u0648 \u0622\u0632\u0645\u0627\u06cc\u0634\u06cc:<\/p>\n<pre><code class=\"hljs\"><span class=\"hljs-keyword\">import<\/span> matplotlib.pyplot <span class=\"hljs-keyword\">as<\/span> plt\n\nplt.plot(history.history(<span class=\"hljs-string\">'acc'<\/span>))\nplt.plot(history.history(<span class=\"hljs-string\">'val_acc'<\/span>))\n\nplt.title(<span class=\"hljs-string\">'model accuracy'<\/span>)\nplt.ylabel(<span class=\"hljs-string\">'accuracy'<\/span>)\nplt.xlabel(<span class=\"hljs-string\">'epoch'<\/span>)\nplt.legend((<span class=\"hljs-string\">'train'<\/span>,<span class=\"hljs-string\">'test'<\/span>), loc=<span class=\"hljs-string\">'upper left'<\/span>)\nplt.show()\n\nplt.plot(history.history(<span class=\"hljs-string\">'loss'<\/span>))\nplt.plot(history.history(<span class=\"hljs-string\">'val_loss'<\/span>))\n\nplt.title(<span class=\"hljs-string\">'model loss'<\/span>)\nplt.ylabel(<span class=\"hljs-string\">'loss'<\/span>)\nplt.xlabel(<span class=\"hljs-string\">'epoch'<\/span>)\nplt.legend((<span class=\"hljs-string\">'train'<\/span>,<span class=\"hljs-string\">'test'<\/span>), loc=<span class=\"hljs-string\">'upper left'<\/span>)\nplt.show()\n<\/code><\/pre>\n<p><img decoding=\"async\" class=\"img-responsive\" src=\"https:\/\/rasanegar.com\/blog\/wp-content\/uploads\/2024\/01\/python-for-nlp-creating-text-classification-model-with-multiple-inputs-in-keras-11.png\" alt=\"\u0633\u0631\" title=\"\"><\/p>\n<p>\u0645\u06cc\u200c\u0628\u06cc\u0646\u06cc\u062f \u06a9\u0647 \u062a\u0641\u0627\u0648\u062a\u200c\u0647\u0627 \u0628\u0631\u0627\u06cc \u0645\u0642\u0627\u062f\u06cc\u0631 \u062a\u0644\u0641\u0627\u062a \u0648 \u062f\u0642\u062a \u0628\u06cc\u0646 \u0645\u062c\u0645\u0648\u0639\u0647\u200c\u0647\u0627\u06cc \u0622\u0645\u0648\u0632\u0634\u06cc \u0648 \u062a\u0633\u062a \u062d\u062f\u0627\u0642\u0644 \u0627\u0633\u062a\u060c \u0628\u0646\u0627\u0628\u0631\u0627\u06cc\u0646 \u0645\u062f\u0644 \u0645\u0627 \u0628\u06cc\u0634 \u0627\u0632 \u062d\u062f \u0645\u0646\u0627\u0633\u0628 \u0646\u06cc\u0633\u062a.<\/p>\n<h2 id=\"finalthoughtsandimprovements\"><span class=\"ez-toc-section\" id=\"%d8%a7%d9%81%da%a9%d8%a7%d8%b1_%d9%86%d9%87%d8%a7%db%8c%db%8c_%d9%88_%d8%a8%d9%87%d8%a8%d9%88%d8%af%d9%87%d8%a7\"><\/span>\u0627\u0641\u06a9\u0627\u0631 \u0646\u0647\u0627\u06cc\u06cc \u0648 \u0628\u0647\u0628\u0648\u062f\u0647\u0627<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<p>\u062f\u0631 \u0627\u06cc\u0646 \u0645\u0642\u0627\u0644\u0647\u060c \u0645\u0627 \u06cc\u06a9 \u0634\u0628\u06a9\u0647 \u0639\u0635\u0628\u06cc \u0628\u0633\u06cc\u0627\u0631 \u0633\u0627\u062f\u0647 \u0633\u0627\u062e\u062a\u06cc\u0645 \u0632\u06cc\u0631\u0627 \u0647\u062f\u0641 \u0645\u0642\u0627\u0644\u0647 \u062a\u0648\u0636\u06cc\u062d \u0686\u06af\u0648\u0646\u06af\u06cc \u0627\u06cc\u062c\u0627\u062f \u06cc\u06a9 \u0645\u062f\u0644 \u06cc\u0627\u062f\u06af\u06cc\u0631\u06cc \u0639\u0645\u06cc\u0642 \u0627\u0633\u062a \u06a9\u0647 \u0648\u0631\u0648\u062f\u06cc \u0647\u0627\u06cc \u0645\u062a\u0639\u062f\u062f \u0627\u0632 \u0627\u0646\u0648\u0627\u0639 \u0645\u062e\u062a\u0644\u0641 \u0631\u0627 \u0645\u06cc \u067e\u0630\u06cc\u0631\u062f.<\/p>\n<p>\u062f\u0631 \u0632\u06cc\u0631 \u0628\u0631\u062e\u06cc \u0627\u0632 \u0646\u06a9\u0627\u062a\u06cc \u0648\u062c\u0648\u062f \u062f\u0627\u0631\u062f \u06a9\u0647 \u0645\u06cc \u062a\u0648\u0627\u0646\u06cc\u062f \u0628\u0631\u0627\u06cc \u0628\u0647\u0628\u0648\u062f \u0639\u0645\u0644\u06a9\u0631\u062f \u0645\u062f\u0644 \u0637\u0628\u0642\u0647 \u0628\u0646\u062f\u06cc \u0645\u062a\u0646 \u062f\u0646\u0628\u0627\u0644 \u06a9\u0646\u06cc\u062f:<\/p>\n<ol>\n<li>\u0645\u0627 \u0641\u0642\u0637 \u0627\u0632 50000 \u0631\u06a9\u0648\u0631\u062f \u0627\u0632 5.2 \u0645\u06cc\u0644\u06cc\u0648\u0646 \u0631\u06a9\u0648\u0631\u062f \u062f\u0631 \u0627\u06cc\u0646 \u0645\u0642\u0627\u0644\u0647 \u0627\u0633\u062a\u0641\u0627\u062f\u0647 \u06a9\u0631\u062f\u06cc\u0645 \u0632\u06cc\u0631\u0627 \u0645\u062d\u062f\u0648\u062f\u06cc\u062a \u0647\u0627\u06cc \u0633\u062e\u062a \u0627\u0641\u0632\u0627\u0631\u06cc \u062f\u0627\u0634\u062a\u06cc\u0645.  \u0645\u06cc \u062a\u0648\u0627\u0646\u06cc\u062f \u0633\u0639\u06cc \u06a9\u0646\u06cc\u062f \u0645\u062f\u0644 \u062e\u0648\u062f \u0631\u0627 \u0622\u0645\u0648\u0632\u0634 \u062f\u0647\u06cc\u062f \u0631\u0648\u06cc \u062a\u0639\u062f\u0627\u062f \u0631\u06a9\u0648\u0631\u062f\u0647\u0627\u06cc \u0628\u0627\u0644\u0627\u062a\u0631\u06cc \u062f\u0627\u0634\u062a\u0647 \u0628\u0627\u0634\u06cc\u062f \u0648 \u0628\u0628\u06cc\u0646\u06cc\u062f \u0622\u06cc\u0627 \u0645\u06cc \u062a\u0648\u0627\u0646\u06cc\u062f \u0639\u0645\u0644\u06a9\u0631\u062f \u0628\u0647\u062a\u0631\u06cc \u062f\u0627\u0634\u062a\u0647 \u0628\u0627\u0634\u06cc\u062f \u06cc\u0627 \u062e\u06cc\u0631.<\/li>\n<li>\u0633\u0639\u06cc \u06a9\u0646\u06cc\u062f LSTM \u0648 \u0644\u0627\u06cc\u0647 \u0647\u0627\u06cc \u0645\u062a\u0631\u0627\u06a9\u0645 \u0628\u06cc\u0634\u062a\u0631\u06cc \u0631\u0627 \u0628\u0647 \u0645\u062f\u0644 \u0627\u0636\u0627\u0641\u0647 \u06a9\u0646\u06cc\u062f.  \u0627\u06af\u0631 \u0645\u062f\u0644 \u0628\u06cc\u0634 \u0627\u0632 \u062d\u062f \u0645\u0646\u0627\u0633\u0628 \u0627\u0633\u062a\u060c \u0633\u0639\u06cc \u06a9\u0646\u06cc\u062f \u062a\u0631\u06a9 \u062a\u062d\u0635\u06cc\u0644 \u0631\u0627 \u0627\u0636\u0627\u0641\u0647 \u06a9\u0646\u06cc\u062f.<\/li>\n<li>\u0633\u0639\u06cc \u06a9\u0646\u06cc\u062f \u0639\u0645\u0644\u06a9\u0631\u062f \u0628\u0647\u06cc\u0646\u0647 \u0633\u0627\u0632 \u0631\u0627 \u062a\u063a\u06cc\u06cc\u0631 \u062f\u0647\u06cc\u062f \u0648 \u0645\u062f\u0644 \u0631\u0627 \u0628\u0627 \u062a\u0639\u062f\u0627\u062f \u062f\u0648\u0631\u0647 \u0647\u0627\u06cc \u0628\u0627\u0644\u0627\u062a\u0631 \u0622\u0645\u0648\u0632\u0634 \u062f\u0647\u06cc\u062f.<\/li>\n<\/ol>\n<p>\u0644\u0637\u0641\u0627 \u0646\u062a\u0627\u06cc\u062c \u062e\u0648\u062f \u0631\u0627 \u0647\u0645\u0631\u0627\u0647 \u0628\u0627 \u067e\u06cc\u06a9\u0631\u0628\u0646\u062f\u06cc \u0634\u0628\u06a9\u0647 \u0639\u0635\u0628\u06cc \u062f\u0631 \u0642\u0633\u0645\u062a \u0646\u0638\u0631\u0627\u062a \u0628\u0647 \u0627\u0634\u062a\u0631\u0627\u06a9 \u0628\u06af\u0630\u0627\u0631\u06cc\u062f.  \u062e\u06cc\u0644\u06cc \u062f\u0648\u0633\u062a \u062f\u0627\u0631\u0645 \u0628\u0628\u06cc\u0646\u0645 \u0686\u0642\u062f\u0631 \u062e\u0648\u0628 \u0627\u062c\u0631\u0627 \u0645\u06cc\u06a9\u0646\u06cc<\/p>\n<\/div>\n<p><script>\n                        !function(f,b,e,v,n,t,s)\n                        {if(f.fbq)return;n=f.fbq=function(){n.callMethod?\n                        n.callMethod.apply(n,arguments):n.queue.push(arguments)};\n                        if(!f._fbq)f._fbq=n;n.push=n;n.loaded=!0;n.version='2.0';\n                        n.queue=();t=b.createElement(e);t.async=!0;\n                        t.src=v;s=b.getElementsByTagName(e)(0);\n                        s.parentNode.insertBefore(t,s)}(window, document,'script',\n                        'https:\/\/connect.facebook.net\/en_US\/fbevents.js');\n                        fbq('init', '525232124909042');\n                        fbq('track', 'PageView');\n                    <\/script>    (\u0628\u0631\u0686\u0633\u0628\u200c\u0647\u0627 \u0628\u0647 \u062a\u0631\u062c\u0645\u0647)# python<br \/>\n<br \/><br \/>\n<br \/>\u0645\u0646\u062a\u0634\u0631 \u0634\u062f\u0647 \u062f\u0631 1403-01-20 20:57:04<br \/>\n<\/p>\n\n\n<div class=\"kk-star-ratings kksr-auto kksr-align-center kksr-valign-bottom\"\n    data-payload='{&quot;align&quot;:&quot;center&quot;,&quot;id&quot;:&quot;16103&quot;,&quot;slug&quot;:&quot;default&quot;,&quot;valign&quot;:&quot;bottom&quot;,&quot;ignore&quot;:&quot;&quot;,&quot;reference&quot;:&quot;auto&quot;,&quot;class&quot;:&quot;&quot;,&quot;count&quot;:&quot;0&quot;,&quot;legendonly&quot;:&quot;&quot;,&quot;readonly&quot;:&quot;&quot;,&quot;score&quot;:&quot;0&quot;,&quot;starsonly&quot;:&quot;&quot;,&quot;best&quot;:&quot;5&quot;,&quot;gap&quot;:&quot;5&quot;,&quot;greet&quot;:&quot;\u0627\u0645\u062a\u06cc\u0627\u0632 \u0634\u0645\u0627 \u0628\u0647 \u0627\u06cc\u0646 \u0645\u0637\u0644\u0628&quot;,&quot;legend&quot;:&quot;0\\\/5 (0 \u0631\u0627\u06cc)&quot;,&quot;size&quot;:&quot;30&quot;,&quot;title&quot;:&quot;Python \u0628\u0631\u0627\u06cc NLP: \u0627\u06cc\u062c\u0627\u062f \u0645\u062f\u0644\u200c\u0647\u0627\u06cc \u0637\u0628\u0642\u0647\u200c\u0628\u0646\u062f\u06cc \u0686\u0646\u062f \u062f\u0627\u062f\u0647 \u0628\u0627 Keras&quot;,&quot;width&quot;:&quot;0&quot;,&quot;_legend&quot;:&quot;{score}\\\/{best} ({count} \u0631\u0627\u06cc)&quot;,&quot;font_factor&quot;:&quot;1.25&quot;}'>\n            \n<div class=\"kksr-stars\">\n    \n<div class=\"kksr-stars-inactive\">\n            <div class=\"kksr-star\" data-star=\"1\" style=\"padding-left: 5px\">\n            \n\n<div class=\"kksr-icon\" style=\"width: 30px; height: 30px;\"><\/div>\n        <\/div>\n            <div class=\"kksr-star\" data-star=\"2\" style=\"padding-left: 5px\">\n            \n\n<div class=\"kksr-icon\" style=\"width: 30px; height: 30px;\"><\/div>\n        <\/div>\n            <div class=\"kksr-star\" data-star=\"3\" style=\"padding-left: 5px\">\n            \n\n<div class=\"kksr-icon\" style=\"width: 30px; height: 30px;\"><\/div>\n        <\/div>\n            <div class=\"kksr-star\" data-star=\"4\" style=\"padding-left: 5px\">\n            \n\n<div class=\"kksr-icon\" style=\"width: 30px; height: 30px;\"><\/div>\n        <\/div>\n            <div class=\"kksr-star\" data-star=\"5\" style=\"padding-left: 5px\">\n            \n\n<div class=\"kksr-icon\" style=\"width: 30px; height: 30px;\"><\/div>\n        <\/div>\n    <\/div>\n    \n<div class=\"kksr-stars-active\" style=\"width: 0px;\">\n            <div class=\"kksr-star\" style=\"padding-left: 5px\">\n            \n\n<div class=\"kksr-icon\" style=\"width: 30px; height: 30px;\"><\/div>\n        <\/div>\n            <div class=\"kksr-star\" style=\"padding-left: 5px\">\n            \n\n<div class=\"kksr-icon\" style=\"width: 30px; height: 30px;\"><\/div>\n        <\/div>\n            <div class=\"kksr-star\" style=\"padding-left: 5px\">\n            \n\n<div class=\"kksr-icon\" style=\"width: 30px; height: 30px;\"><\/div>\n        <\/div>\n            <div class=\"kksr-star\" style=\"padding-left: 5px\">\n            \n\n<div class=\"kksr-icon\" style=\"width: 30px; height: 30px;\"><\/div>\n        <\/div>\n            <div class=\"kksr-star\" style=\"padding-left: 5px\">\n            \n\n<div class=\"kksr-icon\" style=\"width: 30px; height: 30px;\"><\/div>\n        <\/div>\n    <\/div>\n<\/div>\n                \n\n<div class=\"kksr-legend\" style=\"font-size: 24px;\">\n            <span class=\"kksr-muted\">\u0627\u0645\u062a\u06cc\u0627\u0632 \u0634\u0645\u0627 \u0628\u0647 \u0627\u06cc\u0646 \u0645\u0637\u0644\u0628<\/span>\n    <\/div>\n    <\/div>\n","protected":false},"excerpt":{"rendered":"<p><span class=\"span-reading-time rt-reading-time\" style=\"display: block;\"><span class=\"rt-label rt-prefix\">\u0632\u0645\u0627\u0646 \u0644\u0627\u0632\u0645 \u0628\u0631\u0627\u06cc \u0645\u0637\u0627\u0644\u0639\u0647: <\/span> <span class=\"rt-time\"> 17<\/span> <span class=\"rt-label rt-postfix\">\u062f\u0642\u06cc\u0642\u0647<\/span><\/span>\u0627\u06cc\u0646 \u0647\u062c\u062f\u0647\u0645\u06cc\u0646 \u0645\u0642\u0627\u0644\u0647 \u0627\u0632 \u0633\u0631\u06cc \u0645\u0642\u0627\u0644\u0627\u062a \u0645\u0646 \u0627\u0633\u062a \u0631\u0648\u06cc \u067e\u0627\u06cc\u062a\u0648\u0646 \u0628\u0631\u0627\u06cc NLP. \u062f\u0631 \u0645\u0642\u0627\u0644\u0647 \u0642\u0628\u0644\u06cc \u062e\u0648\u062f\u060c \u0631\u0648\u0634 \u0627\u06cc\u062c\u0627\u062f \u06cc\u06a9 \u0645\u062f\u0644 \u062a\u062d\u0644\u06cc\u0644 \u0627\u062d\u0633\u0627\u0633\u0627\u062a \u0641\u06cc\u0644\u0645 \u0645\u0628\u062a\u0646\u06cc \u0628\u0631 \u06cc\u0627\u062f\u06af\u06cc\u0631\u06cc \u0639\u0645\u06cc\u0642 \u0631\u0627 \u0628\u0627 \u0627\u0633\u062a\u0641\u0627\u062f\u0647 \u0627\u0632 \u067e\u0627\u06cc\u062a\u0648\u0646 \u062a\u0648\u0636\u06cc\u062d \u062f\u0627\u062f\u0645. \u06a9\u0631\u0627\u0633 \u06a9\u062a\u0627\u0628\u062e\u0627\u0646\u0647 \u062f\u0631 \u0622\u0646 \u0645\u0642\u0627\u0644\u0647\u060c \u062f\u06cc\u062f\u06cc\u0645 \u06a9\u0647 \u0686\u06af\u0648\u0646\u0647 \u0645\u06cc\u200c\u062a\u0648\u0627\u0646\u06cc\u0645 \u062a\u062d\u0644\u06cc\u0644 \u0627\u062d\u0633\u0627\u0633\u0627\u062a \u0646\u0638\u0631\u0627\u062a \u06a9\u0627\u0631\u0628\u0631\u0627\u0646 \u0631\u0627 \u062f\u0631 \u0645\u0648\u0631\u062f \u0641\u06cc\u0644\u0645\u200c\u0647\u0627\u06cc \u0645\u062e\u062a\u0644\u0641 \u0627\u0646\u062c\u0627\u0645 \u062f\u0647\u06cc\u0645 \u0631\u0648\u06cc [&hellip;]<\/p>\n","protected":false},"author":3,"featured_media":16104,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[1743,620],"tags":[],"class_list":["post-16103","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-python","category-programming"],"acf":[],"_links":{"self":[{"href":"https:\/\/rasanegaar.com\/blog\/wp-json\/wp\/v2\/posts\/16103","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/rasanegaar.com\/blog\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/rasanegaar.com\/blog\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/rasanegaar.com\/blog\/wp-json\/wp\/v2\/users\/3"}],"replies":[{"embeddable":true,"href":"https:\/\/rasanegaar.com\/blog\/wp-json\/wp\/v2\/comments?post=16103"}],"version-history":[{"count":0,"href":"https:\/\/rasanegaar.com\/blog\/wp-json\/wp\/v2\/posts\/16103\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/rasanegaar.com\/blog\/wp-json\/wp\/v2\/media\/16104"}],"wp:attachment":[{"href":"https:\/\/rasanegaar.com\/blog\/wp-json\/wp\/v2\/media?parent=16103"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/rasanegaar.com\/blog\/wp-json\/wp\/v2\/categories?post=16103"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/rasanegaar.com\/blog\/wp-json\/wp\/v2\/tags?post=16103"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}