{"id":16075,"date":"2024-01-20T12:02:28","date_gmt":"2024-01-20T08:32:28","guid":{"rendered":"https:\/\/rasanegar.com\/blog\/%d8%ad%d9%84-%d9%85%d8%b3%d8%a7%d8%a6%d9%84-%d8%aa%d9%88%d8%a7%d9%84%db%8c-%d8%a8%d8%a7-lstm-%d8%af%d8%b1-keras\/"},"modified":"2024-01-20T12:02:28","modified_gmt":"2024-01-20T08:32:28","slug":"%d8%ad%d9%84-%d9%85%d8%b3%d8%a7%d8%a6%d9%84-%d8%aa%d9%88%d8%a7%d9%84%db%8c-%d8%a8%d8%a7-lstm-%d8%af%d8%b1-keras","status":"publish","type":"post","link":"https:\/\/rasanegaar.com\/blog\/%d8%ad%d9%84-%d9%85%d8%b3%d8%a7%d8%a6%d9%84-%d8%aa%d9%88%d8%a7%d9%84%db%8c-%d8%a8%d8%a7-lstm-%d8%af%d8%b1-keras\/","title":{"rendered":"\u062d\u0644 \u0645\u0633\u0627\u0626\u0644 \u062a\u0648\u0627\u0644\u06cc \u0628\u0627 LSTM \u062f\u0631 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\/%d8%ad%d9%84-%d9%85%d8%b3%d8%a7%d8%a6%d9%84-%d8%aa%d9%88%d8%a7%d9%84%db%8c-%d8%a8%d8%a7-lstm-%d8%af%d8%b1-keras\/#%d8%a7%d9%86%d9%88%d8%a7%d8%b9_%d9%85%d8%b3%d8%a7%d8%a6%d9%84_%d8%af%d9%86%d8%a8%d8%a7%d9%84%d9%87_%d8%a7%db%8c\" >\u0627\u0646\u0648\u0627\u0639 \u0645\u0633\u0627\u0626\u0644 \u062f\u0646\u0628\u0627\u0644\u0647 \u0627\u06cc<\/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\/%d8%ad%d9%84-%d9%85%d8%b3%d8%a7%d8%a6%d9%84-%d8%aa%d9%88%d8%a7%d9%84%db%8c-%d8%a8%d8%a7-lstm-%d8%af%d8%b1-keras\/#%d9%85%d8%b4%da%a9%d9%84%d8%a7%d8%aa_%d8%aa%d9%88%d8%a7%d9%84%db%8c_%db%8c%da%a9_%d8%a8%d9%87_%db%8c%da%a9\" >\u0645\u0634\u06a9\u0644\u0627\u062a \u062a\u0648\u0627\u0644\u06cc \u06cc\u06a9 \u0628\u0647 \u06cc\u06a9<\/a><ul class='ez-toc-list-level-3' ><li class='ez-toc-heading-level-3'><a class=\"ez-toc-link ez-toc-heading-3\" href=\"https:\/\/rasanegaar.com\/blog\/%d8%ad%d9%84-%d9%85%d8%b3%d8%a7%d8%a6%d9%84-%d8%aa%d9%88%d8%a7%d9%84%db%8c-%d8%a8%d8%a7-lstm-%d8%af%d8%b1-keras\/#%d9%85%d8%b4%da%a9%d9%84%d8%a7%d8%aa_%d8%aa%d9%88%d8%a7%d9%84%db%8c_%db%8c%da%a9_%d8%a8%d9%87_%db%8c%da%a9_%d8%a8%d8%a7_%db%8c%da%a9_%d9%88%db%8c%da%98%da%af%db%8c_%d9%88%d8%a7%d8%ad%d8%af\" >\u0645\u0634\u06a9\u0644\u0627\u062a \u062a\u0648\u0627\u0644\u06cc \u06cc\u06a9 \u0628\u0647 \u06cc\u06a9 \u0628\u0627 \u06cc\u06a9 \u0648\u06cc\u0698\u06af\u06cc \u0648\u0627\u062d\u062f<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-3'><a class=\"ez-toc-link ez-toc-heading-4\" href=\"https:\/\/rasanegaar.com\/blog\/%d8%ad%d9%84-%d9%85%d8%b3%d8%a7%d8%a6%d9%84-%d8%aa%d9%88%d8%a7%d9%84%db%8c-%d8%a8%d8%a7-lstm-%d8%af%d8%b1-keras\/#%d9%85%d8%b4%da%a9%d9%84%d8%a7%d8%aa_%d8%aa%d9%88%d8%a7%d9%84%db%8c_%db%8c%da%a9_%d8%a8%d9%87_%db%8c%da%a9_%d8%a8%d8%a7_%da%86%d9%86%d8%af%db%8c%d9%86_%d9%88%db%8c%da%98%da%af%db%8c\" >\u0645\u0634\u06a9\u0644\u0627\u062a \u062a\u0648\u0627\u0644\u06cc \u06cc\u06a9 \u0628\u0647 \u06cc\u06a9 \u0628\u0627 \u0686\u0646\u062f\u06cc\u0646 \u0648\u06cc\u0698\u06af\u06cc<\/a><\/li><\/ul><\/li><li class='ez-toc-page-1 ez-toc-heading-level-2'><a class=\"ez-toc-link ez-toc-heading-5\" href=\"https:\/\/rasanegaar.com\/blog\/%d8%ad%d9%84-%d9%85%d8%b3%d8%a7%d8%a6%d9%84-%d8%aa%d9%88%d8%a7%d9%84%db%8c-%d8%a8%d8%a7-lstm-%d8%af%d8%b1-keras\/#%d9%85%d8%b4%da%a9%d9%84%d8%a7%d8%aa_%d8%aa%d9%88%d8%a7%d9%84%db%8c_%da%86%d9%86%d8%af_%d8%a8%d9%87_%db%8c%da%a9\" >\u0645\u0634\u06a9\u0644\u0627\u062a \u062a\u0648\u0627\u0644\u06cc \u0686\u0646\u062f \u0628\u0647 \u06cc\u06a9<\/a><ul class='ez-toc-list-level-3' ><li class='ez-toc-heading-level-3'><a class=\"ez-toc-link ez-toc-heading-6\" href=\"https:\/\/rasanegaar.com\/blog\/%d8%ad%d9%84-%d9%85%d8%b3%d8%a7%d8%a6%d9%84-%d8%aa%d9%88%d8%a7%d9%84%db%8c-%d8%a8%d8%a7-lstm-%d8%af%d8%b1-keras\/#%d9%85%d8%b4%da%a9%d9%84%d8%a7%d8%aa_%d8%aa%d9%88%d8%a7%d9%84%db%8c_%da%86%d9%86%d8%af_%d8%a8%d9%87_%db%8c%da%a9_%d8%a8%d8%a7_%db%8c%da%a9_%d9%88%db%8c%da%98%da%af%db%8c_%d9%88%d8%a7%d8%ad%d8%af\" >\u0645\u0634\u06a9\u0644\u0627\u062a \u062a\u0648\u0627\u0644\u06cc \u0686\u0646\u062f \u0628\u0647 \u06cc\u06a9 \u0628\u0627 \u06cc\u06a9 \u0648\u06cc\u0698\u06af\u06cc \u0648\u0627\u062d\u062f<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-3'><a class=\"ez-toc-link ez-toc-heading-7\" href=\"https:\/\/rasanegaar.com\/blog\/%d8%ad%d9%84-%d9%85%d8%b3%d8%a7%d8%a6%d9%84-%d8%aa%d9%88%d8%a7%d9%84%db%8c-%d8%a8%d8%a7-lstm-%d8%af%d8%b1-keras\/#%d9%85%d8%b4%da%a9%d9%84%d8%a7%d8%aa_%d8%aa%d9%88%d8%a7%d9%84%db%8c_%da%86%d9%86%d8%af_%d8%a8%d9%87_%db%8c%da%a9_%d8%a8%d8%a7_%da%86%d9%86%d8%af%db%8c%d9%86_%d9%88%db%8c%da%98%da%af%db%8c\" >\u0645\u0634\u06a9\u0644\u0627\u062a \u062a\u0648\u0627\u0644\u06cc \u0686\u0646\u062f \u0628\u0647 \u06cc\u06a9 \u0628\u0627 \u0686\u0646\u062f\u06cc\u0646 \u0648\u06cc\u0698\u06af\u06cc<\/a><\/li><\/ul><\/li><li class='ez-toc-page-1 ez-toc-heading-level-2'><a class=\"ez-toc-link ez-toc-heading-8\" href=\"https:\/\/rasanegaar.com\/blog\/%d8%ad%d9%84-%d9%85%d8%b3%d8%a7%d8%a6%d9%84-%d8%aa%d9%88%d8%a7%d9%84%db%8c-%d8%a8%d8%a7-lstm-%d8%af%d8%b1-keras\/#%d9%86%d8%aa%db%8c%d8%ac%d9%87\" >\u0646\u062a\u06cc\u062c\u0647<\/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>\u062f\u0631 \u0627\u06cc\u0646 \u0645\u0642\u0627\u0644\u0647\u060c \u0631\u0648\u0634 \u0627\u0646\u062c\u0627\u0645 \u067e\u06cc\u0634 \u0628\u06cc\u0646\u06cc \u0633\u0631\u06cc \u0647\u0627\u06cc \u0632\u0645\u0627\u0646\u06cc \u0631\u0627 \u06a9\u0647 \u0628\u0631\u0627\u06cc \u062d\u0644 \u0645\u0633\u0627\u0626\u0644 \u062a\u0648\u0627\u0644\u06cc \u0627\u0633\u062a\u0641\u0627\u062f\u0647 \u0645\u06cc \u0634\u0648\u062f\u060c \u06cc\u0627\u062f \u062e\u0648\u0627\u0647\u06cc\u062f \u06af\u0631\u0641\u062a.<\/p>\n<p>\u067e\u06cc\u0634\u200c\u0628\u06cc\u0646\u06cc \u0633\u0631\u06cc\u200c\u0647\u0627\u06cc \u0632\u0645\u0627\u0646\u06cc \u0628\u0647 \u0646\u0648\u0639 \u0645\u0634\u06a9\u0644\u0627\u062a\u06cc \u0627\u0634\u0627\u0631\u0647 \u0645\u06cc\u200c\u06a9\u0646\u062f \u06a9\u0647 \u062f\u0631 \u0622\u0646\u0647\u0627 \u0628\u0627\u06cc\u062f \u06cc\u06a9 \u0646\u062a\u06cc\u062c\u0647 \u0631\u0627 \u067e\u06cc\u0634\u200c\u0628\u06cc\u0646\u06cc \u06a9\u0646\u06cc\u0645 \u0631\u0648\u06cc \u0648\u0631\u0648\u062f\u06cc \u0647\u0627\u06cc \u0648\u0627\u0628\u0633\u062a\u0647 \u0628\u0647 \u0632\u0645\u0627\u0646  \u06cc\u06a9 \u0645\u062b\u0627\u0644 \u0645\u0639\u0645\u0648\u0644\u06cc \u0627\u0632 \u062f\u0627\u062f\u0647 \u0647\u0627\u06cc \u0633\u0631\u06cc \u0632\u0645\u0627\u0646\u06cc\u060c \u062f\u0627\u062f\u0647 \u0647\u0627\u06cc \u0628\u0627\u0632\u0627\u0631 \u0633\u0647\u0627\u0645 \u0627\u0633\u062a \u06a9\u0647 \u062f\u0631 \u0622\u0646 \u0642\u06cc\u0645\u062a \u0633\u0647\u0627\u0645 \u0628\u0627 \u0632\u0645\u0627\u0646 \u062a\u063a\u06cc\u06cc\u0631 \u0645\u06cc \u06a9\u0646\u062f.  \u0628\u0647 \u0647\u0645\u06cc\u0646 \u062a\u0631\u062a\u06cc\u0628\u060c \u062f\u0645\u0627\u06cc \u0633\u0627\u0639\u062a\u06cc \u06cc\u06a9 \u0645\u06a9\u0627\u0646 \u062e\u0627\u0635 \u0646\u06cc\u0632 \u062a\u063a\u06cc\u06cc\u0631 \u0645\u06cc \u06a9\u0646\u062f \u0648 \u0647\u0645\u0686\u0646\u06cc\u0646 \u0645\u06cc \u062a\u0648\u0627\u0646\u062f \u0628\u0647 \u0639\u0646\u0648\u0627\u0646 \u062f\u0627\u062f\u0647 \u0633\u0631\u06cc \u0632\u0645\u0627\u0646\u06cc \u062f\u0631 \u0646\u0638\u0631 \u06af\u0631\u0641\u062a\u0647 \u0634\u0648\u062f.  \u062f\u0627\u062f\u0647\u200c\u0647\u0627\u06cc \u0633\u0631\u06cc \u0632\u0645\u0627\u0646\u06cc \u0627\u0633\u0627\u0633\u0627\u064b \u062f\u0646\u0628\u0627\u0644\u0647\u200c\u0627\u06cc \u0627\u0632 \u062f\u0627\u062f\u0647\u200c\u0647\u0627 \u0647\u0633\u062a\u0646\u062f\u060c \u0627\u0632 \u0627\u06cc\u0646 \u0631\u0648 \u0645\u0633\u0627\u0626\u0644 \u0633\u0631\u06cc \u0632\u0645\u0627\u0646\u06cc \u0627\u063a\u0644\u0628 \u0628\u0647 \u0639\u0646\u0648\u0627\u0646 \u0645\u0633\u0627\u0626\u0644 \u062a\u0648\u0627\u0644\u06cc \u0646\u0627\u0645\u06cc\u062f\u0647 \u0645\u06cc\u200c\u0634\u0648\u0646\u062f.<\/p>\n<p><a rel=\"nofollow noopener\" target=\"_blank\" href=\"https:\/\/en.wikipedia.org\/wiki\/Recurrent_neural_network\">\u0634\u0628\u06a9\u0647 \u0647\u0627\u06cc \u0639\u0635\u0628\u06cc \u0645\u06a9\u0631\u0631<\/a> (RNN) \u062b\u0627\u0628\u062a \u0634\u062f\u0647 \u0627\u0633\u062a \u06a9\u0647 \u0628\u0647 \u0637\u0648\u0631 \u0645\u0648\u062b\u0631 \u0645\u0633\u0627\u0626\u0644 \u062a\u0648\u0627\u0644\u06cc \u0631\u0627 \u062d\u0644 \u0645\u06cc \u06a9\u0646\u062f.  \u0628\u0647 \u0648\u06cc\u0698\u0647\u060c <a rel=\"nofollow noopener\" target=\"_blank\" href=\"https:\/\/en.wikipedia.org\/wiki\/Long_short-term_memory\">\u0634\u0628\u06a9\u0647 \u062d\u0627\u0641\u0638\u0647 \u06a9\u0648\u062a\u0627\u0647 \u0645\u062f\u062a \u0628\u0644\u0646\u062f \u0645\u062f\u062a<\/a> (LSTM)\u060c \u06a9\u0647 \u06cc\u06a9 \u06af\u0648\u0646\u0647 \u0627\u0632 RNN \u0627\u0633\u062a\u060c \u062f\u0631 \u062d\u0627\u0644 \u062d\u0627\u0636\u0631 \u062f\u0631 \u062d\u0648\u0632\u0647 \u0647\u0627\u06cc \u0645\u062e\u062a\u0644\u0641\u06cc \u0628\u0631\u0627\u06cc \u062d\u0644 \u0645\u0633\u0627\u0626\u0644 \u062a\u0648\u0627\u0644\u06cc \u0627\u0633\u062a\u0641\u0627\u062f\u0647 \u0645\u06cc \u0634\u0648\u062f.<\/p>\n<h2 id=\"typesofsequenceproblems\"><span class=\"ez-toc-section\" id=\"%d8%a7%d9%86%d9%88%d8%a7%d8%b9_%d9%85%d8%b3%d8%a7%d8%a6%d9%84_%d8%af%d9%86%d8%a8%d8%a7%d9%84%d9%87_%d8%a7%db%8c\"><\/span>\u0627\u0646\u0648\u0627\u0639 \u0645\u0633\u0627\u0626\u0644 \u062f\u0646\u0628\u0627\u0644\u0647 \u0627\u06cc<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<p>\u0645\u0634\u06a9\u0644\u0627\u062a \u062a\u0648\u0627\u0644\u06cc \u0631\u0627 \u0645\u06cc \u062a\u0648\u0627\u0646 \u0628\u0647 \u0637\u0648\u0631 \u06a9\u0644\u06cc \u0628\u0647 \u062f\u0633\u062a\u0647 \u0647\u0627\u06cc \u0632\u06cc\u0631 \u062f\u0633\u062a\u0647 \u0628\u0646\u062f\u06cc \u06a9\u0631\u062f:<\/p>\n<ol>\n<li><strong>\u06cc\u06a9 \u0628\u0647 \u06cc\u06a9:<\/strong> \u062c\u0627\u06cc\u06cc \u06a9\u0647 \u06cc\u06a9 \u0648\u0631\u0648\u062f\u06cc \u0648 \u06cc\u06a9 \u062e\u0631\u0648\u062c\u06cc \u0648\u062c\u0648\u062f \u062f\u0627\u0631\u062f.  \u0645\u062b\u0627\u0644 \u0645\u0639\u0645\u0648\u0644\u06cc \u0627\u0632 \u0645\u0634\u06a9\u0644 \u062a\u0648\u0627\u0644\u06cc \u06cc\u06a9 \u0628\u0647 \u06cc\u06a9 \u062d\u0627\u0644\u062a\u06cc \u0627\u0633\u062a \u06a9\u0647 \u0634\u0645\u0627 \u06cc\u06a9 \u062a\u0635\u0648\u06cc\u0631 \u062f\u0627\u0631\u06cc\u062f \u0648 \u0645\u06cc \u062e\u0648\u0627\u0647\u06cc\u062f \u06cc\u06a9 \u0628\u0631\u0686\u0633\u0628 \u0648\u0627\u062d\u062f \u0628\u0631\u0627\u06cc \u062a\u0635\u0648\u06cc\u0631 \u067e\u06cc\u0634 \u0628\u06cc\u0646\u06cc \u06a9\u0646\u06cc\u062f.<\/li>\n<li><strong>\u0686\u0646\u062f \u0628\u0647 \u06cc\u06a9:<\/strong> \u062f\u0631 \u0645\u0633\u0627\u0626\u0644 \u062a\u0648\u0627\u0644\u06cc \u0686\u0646\u062f \u0628\u0647 \u06cc\u06a9\u060c \u062f\u0646\u0628\u0627\u0644\u0647 \u0627\u06cc \u0627\u0632 \u062f\u0627\u062f\u0647 \u0647\u0627 \u0631\u0627 \u0628\u0647 \u0639\u0646\u0648\u0627\u0646 \u0648\u0631\u0648\u062f\u06cc \u062f\u0627\u0631\u06cc\u0645 \u0648 \u0628\u0627\u06cc\u062f \u06cc\u06a9 \u062e\u0631\u0648\u062c\u06cc \u0631\u0627 \u067e\u06cc\u0634 \u0628\u06cc\u0646\u06cc \u06a9\u0646\u06cc\u0645.  \u0637\u0628\u0642\u0647 \u0628\u0646\u062f\u06cc \u0645\u062a\u0646 \u06cc\u06a9 \u0645\u062b\u0627\u0644 \u0627\u0635\u0644\u06cc \u0627\u0632 \u0645\u0633\u0627\u0626\u0644 \u062a\u0648\u0627\u0644\u06cc \u0686\u0646\u062f \u0628\u0647 \u06cc\u06a9 \u0627\u0633\u062a \u06a9\u0647 \u062f\u0631 \u0622\u0646 \u06cc\u06a9 \u062f\u0646\u0628\u0627\u0644\u0647 \u0648\u0631\u0648\u062f\u06cc \u0627\u0632 \u06a9\u0644\u0645\u0627\u062a \u062f\u0627\u0631\u06cc\u0645 \u0648 \u0645\u06cc \u062e\u0648\u0627\u0647\u06cc\u0645 \u06cc\u06a9 \u0628\u0631\u0686\u0633\u0628 \u062e\u0631\u0648\u062c\u06cc \u0631\u0627 \u067e\u06cc\u0634 \u0628\u06cc\u0646\u06cc \u06a9\u0646\u06cc\u0645.<\/li>\n<li><strong>\u06cc\u06a9 \u0628\u0647 \u0686\u0646\u062f:<\/strong> \u062f\u0631 \u0645\u0633\u0627\u0626\u0644 \u062a\u0648\u0627\u0644\u06cc \u06cc\u06a9 \u0628\u0647 \u0686\u0646\u062f\u060c \u06cc\u06a9 \u0648\u0631\u0648\u062f\u06cc \u0648\u0627\u062d\u062f \u0648 \u06cc\u06a9 \u062a\u0648\u0627\u0644\u06cc \u062e\u0631\u0648\u062c\u06cc \u062f\u0627\u0631\u06cc\u0645.  \u06cc\u06a9 \u0645\u062b\u0627\u0644 \u0645\u0639\u0645\u0648\u0644\u06cc \u06cc\u06a9 \u062a\u0635\u0648\u06cc\u0631 \u0648 \u062a\u0648\u0636\u06cc\u062d\u0627\u062a \u0645\u0631\u0628\u0648\u0637 \u0628\u0647 \u0622\u0646 \u0627\u0633\u062a.<\/li>\n<li><strong>\u0686\u0646\u062f \u0628\u0647 \u0686\u0646\u062f<\/strong>: \u0645\u0634\u06a9\u0644\u0627\u062a \u062a\u0648\u0627\u0644\u06cc \u0686\u0646\u062f \u0628\u0647 \u0686\u0646\u062f \u0634\u0627\u0645\u0644 \u06cc\u06a9 \u0648\u0631\u0648\u062f\u06cc \u062f\u0646\u0628\u0627\u0644\u0647 \u0648 \u06cc\u06a9 \u062e\u0631\u0648\u062c\u06cc \u062a\u0648\u0627\u0644\u06cc \u0627\u0633\u062a.  \u0628\u0647 \u0639\u0646\u0648\u0627\u0646 \u0645\u062b\u0627\u0644\u060c \u0642\u06cc\u0645\u062a \u0633\u0647\u0627\u0645 7 \u0631\u0648\u0632 \u0628\u0647 \u0639\u0646\u0648\u0627\u0646 \u0648\u0631\u0648\u062f\u06cc \u0648 \u0642\u06cc\u0645\u062a \u0633\u0647\u0627\u0645 \u062f\u0631 7 \u0631\u0648\u0632 \u0622\u06cc\u0646\u062f\u0647 \u0628\u0647 \u0639\u0646\u0648\u0627\u0646 \u062e\u0631\u0648\u062c\u06cc.  \u0686\u062a \u0628\u0627\u062a \u0647\u0627 \u0646\u06cc\u0632 \u0646\u0645\u0648\u0646\u0647 \u0627\u06cc \u0627\u0632 \u0645\u0634\u06a9\u0644\u0627\u062a \u062a\u0648\u0627\u0644\u06cc \u0686\u0646\u062f \u0628\u0647 \u0686\u0646\u062f \u0647\u0633\u062a\u0646\u062f \u06a9\u0647 \u062f\u0631 \u0622\u0646 \u06cc\u06a9 \u062f\u0646\u0628\u0627\u0644\u0647 \u0645\u062a\u0646 \u0648\u0631\u0648\u062f\u06cc \u0648 \u062f\u0646\u0628\u0627\u0644\u0647 \u0645\u062a\u0646\u06cc \u062f\u06cc\u06af\u0631 \u062e\u0631\u0648\u062c\u06cc \u0627\u0633\u062a.<\/li>\n<\/ol>\n<p>\u0627\u06cc\u0646 \u0645\u0642\u0627\u0644\u0647 \u0642\u0633\u0645\u062a 1 \u0627\u0632 \u0645\u062c\u0645\u0648\u0639\u0647 \u0627\u0633\u062a.  \u062f\u0631 \u0627\u06cc\u0646 \u0645\u0642\u0627\u0644\u0647 \u062e\u0648\u0627\u0647\u06cc\u0645 \u062f\u06cc\u062f \u06a9\u0647 \u0686\u06af\u0648\u0646\u0647 \u0645\u06cc \u062a\u0648\u0627\u0646 \u0627\u0632 LSTM \u0648 \u0627\u0646\u0648\u0627\u0639 \u0645\u062e\u062a\u0644\u0641 \u0622\u0646 \u0628\u0631\u0627\u06cc \u062d\u0644 \u0645\u0633\u0627\u0626\u0644 \u062a\u0648\u0627\u0644\u06cc \u06cc\u06a9 \u0628\u0647 \u06cc\u06a9 \u0648 \u0686\u0646\u062f \u0628\u0647 \u06cc\u06a9 \u0627\u0633\u062a\u0641\u0627\u062f\u0647 \u06a9\u0631\u062f.  \u062f\u0631 \u0642\u0633\u0645\u062a \u0628\u0639\u062f\u06cc \u0627\u06cc\u0646 \u0645\u062c\u0645\u0648\u0639\u0647 \u0631\u0648\u0634 \u062d\u0644 \u0645\u0633\u0627\u0626\u0644 \u062a\u0648\u0627\u0644\u06cc \u06cc\u06a9 \u0628\u0647 \u0686\u0646\u062f \u0648 \u0686\u0646\u062f \u0628\u0647 \u0686\u0646\u062f \u0631\u0627 \u062e\u0648\u0627\u0647\u06cc\u0645 \u062f\u06cc\u062f.  \u0645\u0627 \u0628\u0627 \u06a9\u062a\u0627\u0628\u062e\u0627\u0646\u0647 Keras \u067e\u0627\u06cc\u062a\u0648\u0646 \u06a9\u0627\u0631 \u062e\u0648\u0627\u0647\u06cc\u0645 \u06a9\u0631\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 \u062a\u0648\u0627\u0646\u06cc\u062f \u0645\u0634\u06a9\u0644\u0627\u062a\u06cc \u0645\u0627\u0646\u0646\u062f \u067e\u06cc\u0634 \u0628\u06cc\u0646\u06cc \u0642\u06cc\u0645\u062a \u0633\u0647\u0627\u0645\u060c \u067e\u06cc\u0634 \u0628\u06cc\u0646\u06cc \u0622\u0628 \u0648 \u0647\u0648\u0627 \u0648 \u063a\u06cc\u0631\u0647 \u0631\u0627 \u0628\u0631 \u0627\u0633\u0627\u0633 \u062d\u0644 \u06a9\u0646\u06cc\u062f. \u0631\u0648\u06cc \u062f\u0627\u062f\u0647 \u0647\u0627\u06cc \u062a\u0627\u0631\u06cc\u062e\u06cc  \u0627\u0632 \u0622\u0646\u062c\u0627\u06cc\u06cc \u06a9\u0647 \u0645\u062a\u0646 \u0646\u06cc\u0632 \u062f\u0646\u0628\u0627\u0644\u0647 \u0627\u06cc \u0627\u0632 \u06a9\u0644\u0645\u0627\u062a \u0627\u0633\u062a\u060c \u062f\u0627\u0646\u0634 \u0628\u0647 \u062f\u0633\u062a \u0622\u0645\u062f\u0647 \u062f\u0631 \u0627\u06cc\u0646 \u0645\u0642\u0627\u0644\u0647 \u0645\u06cc \u062a\u0648\u0627\u0646\u062f \u0628\u0631\u0627\u06cc \u062d\u0644 \u0648\u0638\u0627\u06cc\u0641 \u067e\u0631\u062f\u0627\u0632\u0634 \u0632\u0628\u0627\u0646 \u0637\u0628\u06cc\u0639\u06cc \u0645\u0627\u0646\u0646\u062f \u0637\u0628\u0642\u0647 \u0628\u0646\u062f\u06cc \u0645\u062a\u0646\u060c \u062a\u0648\u0644\u06cc\u062f \u0632\u0628\u0627\u0646 \u0648 \u063a\u06cc\u0631\u0647 \u0646\u06cc\u0632 \u0645\u0648\u0631\u062f \u0627\u0633\u062a\u0641\u0627\u062f\u0647 \u0642\u0631\u0627\u0631 \u06af\u06cc\u0631\u062f.<\/p>\n<h2 id=\"onetoonesequenceproblems\"><span class=\"ez-toc-section\" id=\"%d9%85%d8%b4%da%a9%d9%84%d8%a7%d8%aa_%d8%aa%d9%88%d8%a7%d9%84%db%8c_%db%8c%da%a9_%d8%a8%d9%87_%db%8c%da%a9\"><\/span>\u0645\u0634\u06a9\u0644\u0627\u062a \u062a\u0648\u0627\u0644\u06cc \u06cc\u06a9 \u0628\u0647 \u06cc\u06a9<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<p>\u0647\u0645\u0627\u0646\u0637\u0648\u0631 \u06a9\u0647 \u0642\u0628\u0644\u0627\u064b \u06af\u0641\u062a\u0645\u060c \u062f\u0631 \u0645\u0633\u0627\u0626\u0644 \u062a\u0648\u0627\u0644\u06cc \u06cc\u06a9 \u0628\u0647 \u06cc\u06a9\u060c \u06cc\u06a9 \u0648\u0631\u0648\u062f\u06cc \u0648 \u06cc\u06a9 \u062e\u0631\u0648\u062c\u06cc \u0648\u062c\u0648\u062f \u062f\u0627\u0631\u062f.  \u062f\u0631 \u0627\u06cc\u0646 \u0628\u062e\u0634 \u062f\u0648 \u0646\u0648\u0639 \u0645\u0634\u06a9\u0644 \u062a\u0648\u0627\u0644\u06cc \u0631\u0627 \u0645\u0634\u0627\u0647\u062f\u0647 \u062e\u0648\u0627\u0647\u06cc\u0645 \u06a9\u0631\u062f.  \u0627\u0628\u062a\u062f\u0627 \u0631\u0648\u0634 \u062d\u0644 \u0645\u0633\u0627\u0626\u0644 \u062a\u0648\u0627\u0644\u06cc \u06cc\u06a9 \u0628\u0647 \u06cc\u06a9 \u0631\u0627 \u0628\u0627 \u06cc\u06a9 \u0648\u06cc\u0698\u06af\u06cc \u0648\u0627\u062d\u062f \u0648 \u0633\u067e\u0633 \u062e\u0648\u0627\u0647\u06cc\u0645 \u062f\u06cc\u062f \u06a9\u0647 \u0686\u06af\u0648\u0646\u0647 \u0645\u0633\u0627\u0626\u0644 \u062a\u0648\u0627\u0644\u06cc \u06cc\u06a9 \u0628\u0647 \u06cc\u06a9 \u0631\u0627 \u0628\u0627 \u0686\u0646\u062f\u06cc\u0646 \u0648\u06cc\u0698\u06af\u06cc \u062d\u0644 \u06a9\u0646\u06cc\u0645.<\/p>\n<h3 id=\"onetoonesequenceproblemswithasinglefeature\"><span class=\"ez-toc-section\" id=\"%d9%85%d8%b4%da%a9%d9%84%d8%a7%d8%aa_%d8%aa%d9%88%d8%a7%d9%84%db%8c_%db%8c%da%a9_%d8%a8%d9%87_%db%8c%da%a9_%d8%a8%d8%a7_%db%8c%da%a9_%d9%88%db%8c%da%98%da%af%db%8c_%d9%88%d8%a7%d8%ad%d8%af\"><\/span>\u0645\u0634\u06a9\u0644\u0627\u062a \u062a\u0648\u0627\u0644\u06cc \u06cc\u06a9 \u0628\u0647 \u06cc\u06a9 \u0628\u0627 \u06cc\u06a9 \u0648\u06cc\u0698\u06af\u06cc \u0648\u0627\u062d\u062f<span class=\"ez-toc-section-end\"><\/span><\/h3>\n<p>\u062f\u0631 \u0627\u06cc\u0646 \u0628\u062e\u0634\u060c \u0631\u0648\u0634 \u062d\u0644 \u06cc\u06a9 \u0645\u0633\u0626\u0644\u0647 \u062a\u0648\u0627\u0644\u06cc \u06cc\u06a9 \u0628\u0647 \u06cc\u06a9 \u0631\u0627 \u062e\u0648\u0627\u0647\u06cc\u0645 \u062f\u06cc\u062f \u06a9\u0647 \u0647\u0631 \u0645\u0631\u062d\u0644\u0647 \u0632\u0645\u0627\u0646\u06cc \u062f\u0627\u0631\u0627\u06cc \u06cc\u06a9 \u0648\u06cc\u0698\u06af\u06cc \u0648\u0627\u062d\u062f \u0627\u0633\u062a.<\/p>\n<p>\u0627\u0648\u0644 \u0628\u06cc\u0627\u06cc\u06cc\u062f import \u06a9\u062a\u0627\u0628\u062e\u0627\u0646\u0647 \u0647\u0627\u06cc \u0645\u0648\u0631\u062f \u0646\u06cc\u0627\u0632\u06cc \u06a9\u0647 \u062f\u0631 \u0627\u06cc\u0646 \u0645\u0642\u0627\u0644\u0647 \u0627\u0632 \u0622\u0646\u0647\u0627 \u0627\u0633\u062a\u0641\u0627\u062f\u0647 \u0645\u06cc \u06a9\u0646\u06cc\u0645:<\/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<span class=\"hljs-keyword\">from<\/span> keras.layers <span class=\"hljs-keyword\">import<\/span> Bidirectional\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\n<span class=\"hljs-keyword\">import<\/span> matplotlib.pyplot <span class=\"hljs-keyword\">as<\/span> plt\n<\/code><\/pre>\n<h4 id=\"creatingthedataset\">\u0627\u06cc\u062c\u0627\u062f \u0645\u062c\u0645\u0648\u0639\u0647 \u062f\u0627\u062f\u0647<\/h4>\n<p>\u062f\u0631 \u0627\u06cc\u0646 \u0645\u0631\u062d\u0644\u0647 \u0628\u0639\u062f\u06cc \u0645\u062c\u0645\u0648\u0639\u0647 \u062f\u0627\u062f\u0647 \u0627\u06cc \u0631\u0627 \u06a9\u0647 \u0642\u0631\u0627\u0631 \u0627\u0633\u062a \u0628\u0631\u0627\u06cc \u0627\u06cc\u0646 \u0628\u062e\u0634 \u0627\u0633\u062a\u0641\u0627\u062f\u0647 \u06a9\u0646\u06cc\u0645 \u0622\u0645\u0627\u062f\u0647 \u0645\u06cc \u06a9\u0646\u06cc\u0645.<\/p>\n<pre><code class=\"hljs\">X = <span class=\"hljs-built_in\">list<\/span>()\nY = <span class=\"hljs-built_in\">list<\/span>()\nX = (x+<span class=\"hljs-number\">1<\/span> <span class=\"hljs-keyword\">for<\/span> x <span class=\"hljs-keyword\">in<\/span> <span class=\"hljs-built_in\">range<\/span>(<span class=\"hljs-number\">20<\/span>))\nY = (y * <span class=\"hljs-number\">15<\/span> <span class=\"hljs-keyword\">for<\/span> y <span class=\"hljs-keyword\">in<\/span> X)\n\n<span class=\"hljs-built_in\">print<\/span>(X)\n<span class=\"hljs-built_in\">print<\/span>(Y)\n<\/code><\/pre>\n<p>\u062f\u0631 \u0627\u0633\u06a9\u0631\u06cc\u067e\u062a \u0628\u0627\u0644\u0627 20 \u0648\u0631\u0648\u062f\u06cc \u0648 20 \u062e\u0631\u0648\u062c\u06cc \u0627\u06cc\u062c\u0627\u062f \u0645\u06cc \u06a9\u0646\u06cc\u0645.  \u0647\u0631 \u0648\u0631\u0648\u062f\u06cc \u0634\u0627\u0645\u0644 \u06cc\u06a9 \u0645\u0631\u062d\u0644\u0647 \u0632\u0645\u0627\u0646\u06cc \u0627\u0633\u062a \u06a9\u0647 \u0628\u0647 \u0646\u0648\u0628\u0647 \u062e\u0648\u062f \u0634\u0627\u0645\u0644 \u06cc\u06a9 \u0648\u06cc\u0698\u06af\u06cc \u0648\u0627\u062d\u062f \u0627\u0633\u062a.  \u0647\u0631 \u0645\u0642\u062f\u0627\u0631 \u062e\u0631\u0648\u062c\u06cc \u0627\u0633\u062a <em>15 \u0628\u0631\u0627\u0628\u0631 \u0645\u0642\u062f\u0627\u0631 \u0648\u0631\u0648\u062f\u06cc \u0645\u0631\u0628\u0648\u0637\u0647<\/em>.  \u0627\u06af\u0631 \u0627\u0633\u06a9\u0631\u06cc\u067e\u062a \u0628\u0627\u0644\u0627 \u0631\u0627 \u0627\u062c\u0631\u0627 \u06a9\u0646\u06cc\u062f\u060c \u0628\u0627\u06cc\u062f \u0645\u0642\u0627\u062f\u06cc\u0631 \u0648\u0631\u0648\u062f\u06cc \u0648 \u062e\u0631\u0648\u062c\u06cc \u0631\u0627 \u0645\u0637\u0627\u0628\u0642 \u0634\u06a9\u0644 \u0632\u06cc\u0631 \u0645\u0634\u0627\u0647\u062f\u0647 \u06a9\u0646\u06cc\u062f:<\/p>\n<pre><code class=\"hljs\">(1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20)\n(15, 30, 45, 60, 75, 90, 105, 120, 135, 150, 165, 180, 195, 210, 225, 240, 255, 270, 285, 300)\n<\/code><\/pre>\n<p>\u0648\u0631\u0648\u062f\u06cc \u0644\u0627\u06cc\u0647 LSTM \u0628\u0627\u06cc\u062f \u0628\u0647 \u0634\u06a9\u0644 \u0633\u0647 \u0628\u0639\u062f\u06cc \u0628\u0627\u0634\u062f (\u0646\u0645\u0648\u0646\u0647 \u0647\u0627\u060c \u0645\u0631\u0627\u062d\u0644 \u0632\u0645\u0627\u0646\u06cc\u060c \u0648\u06cc\u0698\u06af\u06cc \u0647\u0627).  \u0646\u0645\u0648\u0646\u0647 \u0647\u0627 \u062a\u0639\u062f\u0627\u062f \u0646\u0645\u0648\u0646\u0647 \u0647\u0627 \u062f\u0631 \u062f\u0627\u062f\u0647 \u0647\u0627\u06cc \u0648\u0631\u0648\u062f\u06cc \u0627\u0633\u062a.  \u0645\u0627 20 \u0646\u0645\u0648\u0646\u0647 \u062f\u0631 \u0648\u0631\u0648\u062f\u06cc \u062f\u0627\u0631\u06cc\u0645.  \u06af\u0627\u0645 \u0647\u0627\u06cc \u0632\u0645\u0627\u0646\u06cc \u062a\u0639\u062f\u0627\u062f \u06af\u0627\u0645 \u0647\u0627\u06cc \u0632\u0645\u0627\u0646\u06cc \u062f\u0631 \u0647\u0631 \u0646\u0645\u0648\u0646\u0647 \u0627\u0633\u062a.  \u0645\u0627 1 \u0645\u0631\u062d\u0644\u0647 \u0632\u0645\u0627\u0646\u06cc \u062f\u0627\u0631\u06cc\u0645.  \u062f\u0631 \u0646\u0647\u0627\u06cc\u062a\u060c \u0648\u06cc\u0698\u06af\u06cc \u0647\u0627 \u0628\u0627 \u062a\u0639\u062f\u0627\u062f \u0648\u06cc\u0698\u06af\u06cc \u0647\u0627 \u062f\u0631 \u0647\u0631 \u0645\u0631\u062d\u0644\u0647 \u0632\u0645\u0627\u0646\u06cc \u0645\u0637\u0627\u0628\u0642\u062a \u062f\u0627\u0631\u0646\u062f.  \u0645\u0627 \u06cc\u06a9 \u0648\u06cc\u0698\u06af\u06cc \u062f\u0631 \u0647\u0631 \u0645\u0631\u062d\u0644\u0647 \u0632\u0645\u0627\u0646\u06cc \u062f\u0627\u0631\u06cc\u0645.<\/p>\n<p>\u0645\u06cc \u062a\u0648\u0627\u0646\u06cc\u0645 \u062f\u0627\u062f\u0647 \u0647\u0627\u06cc \u062e\u0648\u062f \u0631\u0627 \u0627\u0632 \u0637\u0631\u06cc\u0642 \u062f\u0633\u062a\u0648\u0631 \u0632\u06cc\u0631 \u062a\u063a\u06cc\u06cc\u0631 \u0634\u06a9\u0644 \u062f\u0647\u06cc\u0645:<\/p>\n<pre><code class=\"hljs\">X = array(X).reshape(<span class=\"hljs-number\">20<\/span>, <span class=\"hljs-number\">1<\/span>, <span class=\"hljs-number\">1<\/span>)\n<\/code><\/pre>\n<h4 id=\"solutionviasimplelstm\">\u0631\u0627\u0647 \u062d\u0644 \u0627\u0632 \u0637\u0631\u06cc\u0642 LSTM \u0633\u0627\u062f\u0647<\/h4>\n<p>\u0627\u06a9\u0646\u0648\u0646 \u0645\u06cc \u062a\u0648\u0627\u0646\u06cc\u0645 \u0645\u062f\u0644 \u0633\u0627\u062f\u0647 LSTM \u062e\u0648\u062f \u0631\u0627 \u0628\u0627 \u06cc\u06a9 \u0644\u0627\u06cc\u0647 LSTM \u0627\u06cc\u062c\u0627\u062f \u06a9\u0646\u06cc\u0645.<\/p>\n<pre><code class=\"hljs\">model = Sequential()\nmodel.add(LSTM(<span class=\"hljs-number\">50<\/span>, activation=<span class=\"hljs-string\">'relu'<\/span>, input_shape=(<span class=\"hljs-number\">1<\/span>, <span class=\"hljs-number\">1<\/span>)))\nmodel.add(Dense(<span class=\"hljs-number\">1<\/span>))\nmodel.<span class=\"hljs-built_in\">compile<\/span>(optimizer=<span class=\"hljs-string\">'adam'<\/span>, loss=<span class=\"hljs-string\">'mse'<\/span>)\n<span class=\"hljs-built_in\">print<\/span>(model.summary())\n<\/code><\/pre>\n<p>\u062f\u0631 \u0627\u0633\u06a9\u0631\u06cc\u067e\u062a \u0628\u0627\u0644\u0627\u060c \u06cc\u06a9 \u0645\u062f\u0644 LSTM \u0628\u0627 \u06cc\u06a9 \u0644\u0627\u06cc\u0647 LSTM \u0627\u0632 50 \u0646\u0648\u0631\u0648\u0646 \u0648 <code>relu<\/code> \u062a\u0648\u0627\u0628\u0639 \u0641\u0639\u0627\u0644 \u0633\u0627\u0632\u06cc  \u0645\u06cc \u062a\u0648\u0627\u0646\u06cc\u062f \u0628\u0628\u06cc\u0646\u06cc\u062f \u06a9\u0647 \u0634\u06a9\u0644 \u0648\u0631\u0648\u062f\u06cc (1\u060c1) \u0627\u0633\u062a \u0632\u06cc\u0631\u0627 \u062f\u0627\u062f\u0647 \u0647\u0627\u06cc \u0645\u0627 \u062f\u0627\u0631\u0627\u06cc \u06cc\u06a9 \u0645\u0631\u062d\u0644\u0647 \u0632\u0645\u0627\u0646\u06cc \u0628\u0627 \u06cc\u06a9 \u0648\u06cc\u0698\u06af\u06cc \u0647\u0633\u062a\u0646\u062f.  \u0627\u062c\u0631\u0627\u06cc \u0627\u0633\u06a9\u0631\u06cc\u067e\u062a \u0628\u0627\u0644\u0627 \u062e\u0644\u0627\u0635\u0647 \u0632\u06cc\u0631 \u0631\u0627 \u0686\u0627\u067e \u0645\u06cc \u06a9\u0646\u062f:<\/p>\n<pre><code class=\"hljs\">Layer (type)                 Output Shape              Param #\n=================================================================\nlstm_16 (LSTM)               (None, 50)                10400\n_________________________________________________________________\ndense_15 (Dense)             (None, 1)                 51\n=================================================================\nTotal params: 10,451\nTrainable params: 10,451\nNon-trainable params: 0\n<\/code><\/pre>\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\">model.fit(X, Y, epochs=<span class=\"hljs-number\">2000<\/span>, validation_split=<span class=\"hljs-number\">0.2<\/span>, batch_size=<span class=\"hljs-number\">5<\/span>)\n<\/code><\/pre>\n<p>\u0645\u0627 \u0645\u062f\u0644 \u062e\u0648\u062f \u0631\u0627 \u0628\u0631\u0627\u06cc 2000 \u062f\u0648\u0631\u0647 \u0628\u0627 \u0627\u0646\u062f\u0627\u0632\u0647 \u062f\u0633\u062a\u0647 5 \u0622\u0645\u0648\u0632\u0634 \u0645\u06cc \u062f\u0647\u06cc\u0645. \u0634\u0645\u0627 \u0645\u06cc \u062a\u0648\u0627\u0646\u06cc\u062f \u0647\u0631 \u0639\u062f\u062f\u06cc \u0631\u0627 \u0627\u0646\u062a\u062e\u0627\u0628 \u06a9\u0646\u06cc\u062f.  \u0632\u0645\u0627\u0646\u06cc \u06a9\u0647 \u0645\u062f\u0644 \u0622\u0645\u0648\u0632\u0634 \u062f\u0627\u062f\u0647 \u0634\u062f\u060c \u0645\u06cc\u200c\u062a\u0648\u0627\u0646\u06cc\u0645 \u067e\u06cc\u0634\u200c\u0628\u06cc\u0646\u06cc \u06a9\u0646\u06cc\u0645 \u0631\u0648\u06cc \u06cc\u06a9 \u0646\u0645\u0648\u0646\u0647 \u062c\u062f\u06cc\u062f<\/p>\n<p>\u0641\u0631\u0636 \u06a9\u0646\u06cc\u062f \u0645\u06cc \u062e\u0648\u0627\u0647\u06cc\u0645 \u062e\u0631\u0648\u062c\u06cc \u0631\u0627 \u0628\u0631\u0627\u06cc \u0648\u0631\u0648\u062f\u06cc 30 \u067e\u06cc\u0634 \u0628\u06cc\u0646\u06cc \u06a9\u0646\u06cc\u0645. \u062e\u0631\u0648\u062c\u06cc \u0648\u0627\u0642\u0639\u06cc \u0628\u0627\u06cc\u062f 30 x 15 = 450 \u0628\u0627\u0634\u062f. \u0628\u06cc\u0627\u06cc\u06cc\u062f \u0628\u0628\u06cc\u0646\u06cc\u0645 \u0686\u0647 \u0645\u0642\u062f\u0627\u0631\u06cc \u0628\u0647 \u062f\u0633\u062a \u0645\u06cc \u0622\u0648\u0631\u06cc\u0645.  \u0627\u0628\u062a\u062f\u0627 \u0628\u0627\u06cc\u062f \u062f\u0627\u062f\u0647 \u0647\u0627\u06cc \u0622\u0632\u0645\u0627\u06cc\u0634\u06cc \u062e\u0648\u062f \u0631\u0627 \u0628\u0647 \u0634\u06a9\u0644 \u0645\u0646\u0627\u0633\u0628 \u06cc\u0639\u0646\u06cc \u0634\u06a9\u0644 \u0633\u0647 \u0628\u0639\u062f\u06cc\u060c \u0647\u0645\u0627\u0646\u0637\u0648\u0631 \u06a9\u0647 \u062a\u0648\u0633\u0637 LSTM \u0627\u0646\u062a\u0638\u0627\u0631 \u0645\u06cc \u0631\u0648\u062f\u060c \u062a\u0628\u062f\u06cc\u0644 \u06a9\u0646\u06cc\u0645.  \u0627\u0633\u06a9\u0631\u06cc\u067e\u062a \u0632\u06cc\u0631 \u062e\u0631\u0648\u062c\u06cc \u0639\u062f\u062f 30 \u0631\u0627 \u067e\u06cc\u0634 \u0628\u06cc\u0646\u06cc \u0645\u06cc \u06a9\u0646\u062f:<\/p>\n<pre><code class=\"hljs\">test_input = array((<span class=\"hljs-number\">30<\/span>))\ntest_input = test_input.reshape((<span class=\"hljs-number\">1<\/span>, <span class=\"hljs-number\">1<\/span>, <span class=\"hljs-number\">1<\/span>))\ntest_output = model.predict(test_input, verbose=<span class=\"hljs-number\">0<\/span>)\n<span class=\"hljs-built_in\">print<\/span>(test_output)\n<\/code><\/pre>\n<p>\u0645\u0646 \u06cc\u06a9 \u0645\u0642\u062f\u0627\u0631 \u062e\u0631\u0648\u062c\u06cc \u06af\u0631\u0641\u062a\u0645 <code>437.86<\/code> \u06a9\u0647 \u06a9\u0645\u06cc \u06a9\u0645\u062a\u0631 \u0627\u0632 450 \u0627\u0633\u062a.<\/p>\n<p><strong>\u062a\u0648\u062c\u0647 \u062f\u0627\u0634\u062a\u0647 \u0628\u0627\u0634\u06cc\u062f:<\/strong> \u0630\u06a9\u0631 \u0627\u06cc\u0646 \u0646\u06a9\u062a\u0647 \u0636\u0631\u0648\u0631\u06cc \u0627\u0633\u062a \u06a9\u0647 \u062e\u0631\u0648\u062c\u06cc \u0647\u0627\u06cc\u06cc \u06a9\u0647 \u0628\u0627 \u0627\u062c\u0631\u0627\u06cc \u0627\u0633\u06a9\u0631\u06cc\u067e\u062a \u0647\u0627 \u0628\u0647 \u062f\u0633\u062a \u0645\u06cc \u0622\u0648\u0631\u06cc\u062f \u0628\u0627 \u0645\u0646 \u0645\u062a\u0641\u0627\u0648\u062a \u062e\u0648\u0627\u0647\u062f \u0628\u0648\u062f.  \u0627\u06cc\u0646 \u0628\u0647 \u0627\u06cc\u0646 \u062f\u0644\u06cc\u0644 \u0627\u0633\u062a \u06a9\u0647 \u0634\u0628\u06a9\u0647 \u0639\u0635\u0628\u06cc LSTM \u0648\u0632\u0646 \u0647\u0627 \u0631\u0627 \u0628\u0627 \u0645\u0642\u0627\u062f\u06cc\u0631 \u062a\u0635\u0627\u062f\u0641\u06cc \u0648 \u0645\u0642\u0627\u062f\u06cc\u0631 \u0634\u0645\u0627 \u0645\u0642\u062f\u0627\u0631\u062f\u0647\u06cc \u0627\u0648\u0644\u06cc\u0647 \u0645\u06cc \u06a9\u0646\u062f.  \u0627\u0645\u0627 \u0628\u0647 \u0637\u0648\u0631 \u06a9\u0644\u06cc\u060c \u0646\u062a\u0627\u06cc\u062c \u0646\u0628\u0627\u06cc\u062f \u062a\u0641\u0627\u0648\u062a \u0632\u06cc\u0627\u062f\u06cc \u062f\u0627\u0634\u062a\u0647 \u0628\u0627\u0634\u062f.<\/p>\n<h4 id=\"solutionviastackedlstm\">\u0631\u0627\u0647 \u062d\u0644 \u0627\u0632 \u0637\u0631\u06cc\u0642 Stacked LSTM<\/h4>\n<p>\u0628\u06cc\u0627\u06cc\u06cc\u062f \u0627\u06a9\u0646\u0648\u0646 \u06cc\u06a9 LSTM \u067e\u0634\u062a\u0647\u200c\u0627\u06cc \u0627\u06cc\u062c\u0627\u062f \u06a9\u0646\u06cc\u0645 \u0648 \u0628\u0628\u06cc\u0646\u06cc\u0645 \u0622\u06cc\u0627 \u0645\u06cc\u200c\u062a\u0648\u0627\u0646\u06cc\u0645 \u0646\u062a\u0627\u06cc\u062c \u0628\u0647\u062a\u0631\u06cc \u0628\u06af\u06cc\u0631\u06cc\u0645.  \u0645\u062c\u0645\u0648\u0639\u0647 \u062f\u0627\u062f\u0647 \u062b\u0627\u0628\u062a \u062e\u0648\u0627\u0647\u062f \u0645\u0627\u0646\u062f\u060c \u0645\u062f\u0644 \u062a\u063a\u06cc\u06cc\u0631 \u062e\u0648\u0627\u0647\u062f \u06a9\u0631\u062f.  \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\">model = Sequential()\nmodel.add(LSTM(<span class=\"hljs-number\">50<\/span>, activation=<span class=\"hljs-string\">'relu'<\/span>, return_sequences=<span class=\"hljs-literal\">True<\/span>, input_shape=(<span class=\"hljs-number\">1<\/span>, <span class=\"hljs-number\">1<\/span>)))\nmodel.add(LSTM(<span class=\"hljs-number\">50<\/span>, activation=<span class=\"hljs-string\">'relu'<\/span>))\nmodel.add(Dense(<span class=\"hljs-number\">1<\/span>))\nmodel.<span class=\"hljs-built_in\">compile<\/span>(optimizer=<span class=\"hljs-string\">'adam'<\/span>, loss=<span class=\"hljs-string\">'mse'<\/span>)\n<span class=\"hljs-built_in\">print<\/span>(model.summary())\n<\/code><\/pre>\n<p>\u062f\u0631 \u0645\u062f\u0644 \u0641\u0648\u0642 \u062f\u0648 \u0644\u0627\u06cc\u0647 LSTM \u062f\u0627\u0631\u06cc\u0645.  \u062a\u0648\u062c\u0647 \u06a9\u0646\u06cc\u062f\u060c \u0627\u0648\u0644\u06cc\u0646 \u0644\u0627\u06cc\u0647 LSTM \u062f\u0627\u0631\u0627\u06cc \u067e\u0627\u0631\u0627\u0645\u062a\u0631 \u0627\u0633\u062a <code>return_sequences<\/code>\u060c \u06a9\u0647 \u062a\u0646\u0638\u06cc\u0645 \u0634\u062f\u0647 \u0627\u0633\u062a <code>True<\/code>.  \u0647\u0646\u06af\u0627\u0645\u06cc \u06a9\u0647 \u062f\u0646\u0628\u0627\u0644\u0647 \u0628\u0627\u0632\u06af\u0634\u062a \u0628\u0647 <code>True<\/code>\u060c \u062e\u0631\u0648\u062c\u06cc \u062d\u0627\u0644\u062a \u067e\u0646\u0647\u0627\u0646 \u0647\u0631 \u0646\u0648\u0631\u0648\u0646 \u0628\u0647 \u0639\u0646\u0648\u0627\u0646 \u0648\u0631\u0648\u062f\u06cc \u0644\u0627\u06cc\u0647 LSTM \u0628\u0639\u062f\u06cc \u0627\u0633\u062a\u0641\u0627\u062f\u0647 \u0645\u06cc \u0634\u0648\u062f.  \u062e\u0644\u0627\u0635\u0647 \u0645\u062f\u0644 \u0641\u0648\u0642 \u0628\u0647 \u0634\u0631\u062d \u0632\u06cc\u0631 \u0627\u0633\u062a:<\/p>\n<pre><code class=\"hljs\">_________________________________________________________________\nLayer (type)                 Output Shape              Param #\n=================================================================\nlstm_33 (LSTM)               (None, 1, 50)             10400\n_________________________________________________________________\nlstm_34 (LSTM)               (None, 50)                20200\n_________________________________________________________________\ndense_24 (Dense)             (None, 1)                 51\n=================================================================\nTotal params: 30,651\nTrainable params: 30,651\nNon-trainable params: 0\n________________________\n<\/code><\/pre>\n<p>\u062f\u0631 \u0645\u0631\u062d\u0644\u0647 \u0628\u0639\u062f\u060c \u0628\u0627\u06cc\u062f \u0645\u062f\u0644 \u062e\u0648\u062f \u0631\u0627 \u0647\u0645\u0627\u0646\u0637\u0648\u0631 \u06a9\u0647 \u062f\u0631 \u0627\u0633\u06a9\u0631\u06cc\u067e\u062a \u0632\u06cc\u0631 \u0646\u0634\u0627\u0646 \u062f\u0627\u062f\u0647 \u0634\u062f\u0647 \u0627\u0633\u062a \u0622\u0645\u0648\u0632\u0634 \u062f\u0647\u06cc\u0645:<\/p>\n<pre><code class=\"hljs\">history = model.fit(X, Y, epochs=<span class=\"hljs-number\">2000<\/span>, validation_split=<span class=\"hljs-number\">0.2<\/span>, verbose=<span class=\"hljs-number\">1<\/span>, batch_size=<span class=\"hljs-number\">5<\/span>)\n<\/code><\/pre>\n<p>\u067e\u0633 \u0627\u0632 \u0622\u0645\u0648\u0632\u0634 \u0645\u062f\u0644\u060c \u062f\u0648\u0628\u0627\u0631\u0647 \u067e\u06cc\u0634 \u0628\u06cc\u0646\u06cc \u062e\u0648\u0627\u0647\u06cc\u0645 \u06a9\u0631\u062f \u0631\u0648\u06cc \u0646\u0642\u0637\u0647 \u062f\u0627\u062f\u0647 \u062a\u0633\u062a \u06cc\u0639\u0646\u06cc 30.<\/p>\n<pre><code class=\"hljs\">test_input = array((<span class=\"hljs-number\">30<\/span>))\ntest_input = test_input.reshape((<span class=\"hljs-number\">1<\/span>, <span class=\"hljs-number\">1<\/span>, <span class=\"hljs-number\">1<\/span>))\ntest_output = model.predict(test_input, verbose=<span class=\"hljs-number\">0<\/span>)\n<span class=\"hljs-built_in\">print<\/span>(test_output)\n<\/code><\/pre>\n<p>\u0645\u0646 \u062e\u0631\u0648\u062c\u06cc 459.85 \u06af\u0631\u0641\u062a\u0645 \u06a9\u0647 \u0628\u0647\u062a\u0631 \u0627\u0632 437 \u0627\u0633\u062a\u060c \u0639\u062f\u062f\u06cc \u06a9\u0647 \u0627\u0632 \u0637\u0631\u06cc\u0642 \u06cc\u06a9 \u0644\u0627\u06cc\u0647 LSTM \u0628\u0647 \u062f\u0633\u062a \u0622\u0648\u0631\u062f\u06cc\u0645.<\/p>\n<h3 id=\"onetoonesequenceproblemswithmultiplefeatures\"><span class=\"ez-toc-section\" id=\"%d9%85%d8%b4%da%a9%d9%84%d8%a7%d8%aa_%d8%aa%d9%88%d8%a7%d9%84%db%8c_%db%8c%da%a9_%d8%a8%d9%87_%db%8c%da%a9_%d8%a8%d8%a7_%da%86%d9%86%d8%af%db%8c%d9%86_%d9%88%db%8c%da%98%da%af%db%8c\"><\/span>\u0645\u0634\u06a9\u0644\u0627\u062a \u062a\u0648\u0627\u0644\u06cc \u06cc\u06a9 \u0628\u0647 \u06cc\u06a9 \u0628\u0627 \u0686\u0646\u062f\u06cc\u0646 \u0648\u06cc\u0698\u06af\u06cc<span class=\"ez-toc-section-end\"><\/span><\/h3>\n<p>\u062f\u0631 \u0628\u062e\u0634 \u0622\u062e\u0631\u060c \u0647\u0631 \u0646\u0645\u0648\u0646\u0647 \u0648\u0631\u0648\u062f\u06cc \u062f\u0627\u0631\u0627\u06cc \u06cc\u06a9 \u06af\u0627\u0645 \u0632\u0645\u0627\u0646\u06cc \u0628\u0648\u062f \u06a9\u0647 \u0647\u0631 \u0645\u0631\u062d\u0644\u0647 \u0632\u0645\u0627\u0646\u06cc \u062f\u0627\u0631\u0627\u06cc \u06cc\u06a9 \u0648\u06cc\u0698\u06af\u06cc \u0628\u0648\u062f.  \u062f\u0631 \u0627\u06cc\u0646 \u0628\u062e\u0634 \u062e\u0648\u0627\u0647\u06cc\u0645 \u062f\u06cc\u062f \u06a9\u0647 \u0686\u06af\u0648\u0646\u0647 \u0645\u06cc\u200c\u062a\u0648\u0627\u0646 \u0645\u0633\u0627\u0626\u0644 \u062a\u0648\u0627\u0644\u06cc \u06cc\u06a9 \u0628\u0647 \u06cc\u06a9 \u0631\u0627 \u06a9\u0647 \u062f\u0631 \u0622\u0646 \u06af\u0627\u0645\u200c\u0647\u0627\u06cc \u0632\u0645\u0627\u0646\u06cc \u0648\u0631\u0648\u062f\u06cc \u062f\u0627\u0631\u0627\u06cc \u0686\u0646\u062f\u06cc\u0646 \u0648\u06cc\u0698\u06af\u06cc \u0647\u0633\u062a\u0646\u062f\u060c \u062d\u0644 \u06a9\u0631\u062f.<\/p>\n<h4 id=\"creatingthedataset\">\u0627\u06cc\u062c\u0627\u062f \u0645\u062c\u0645\u0648\u0639\u0647 \u062f\u0627\u062f\u0647<\/h4>\n<p>\u0628\u06cc\u0627\u06cc\u06cc\u062f \u0627\u0628\u062a\u062f\u0627 \u0645\u062c\u0645\u0648\u0639\u0647 \u062f\u0627\u062f\u0647 \u062e\u0648\u062f \u0631\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\">nums = <span class=\"hljs-number\">25<\/span>\n\nX1 = <span class=\"hljs-built_in\">list<\/span>()\nX2 = <span class=\"hljs-built_in\">list<\/span>()\nX = <span class=\"hljs-built_in\">list<\/span>()\nY = <span class=\"hljs-built_in\">list<\/span>()\n\nX1 = ((x+<span class=\"hljs-number\">1<\/span>)*<span class=\"hljs-number\">2<\/span> <span class=\"hljs-keyword\">for<\/span> x <span class=\"hljs-keyword\">in<\/span> <span class=\"hljs-built_in\">range<\/span>(<span class=\"hljs-number\">25<\/span>))\nX2 = ((x+<span class=\"hljs-number\">1<\/span>)*<span class=\"hljs-number\">3<\/span> <span class=\"hljs-keyword\">for<\/span> x <span class=\"hljs-keyword\">in<\/span> <span class=\"hljs-built_in\">range<\/span>(<span class=\"hljs-number\">25<\/span>))\nY = (x1*x2 <span class=\"hljs-keyword\">for<\/span> x1,x2 <span class=\"hljs-keyword\">in<\/span> <span class=\"hljs-built_in\">zip<\/span>(X1,X2))\n\n<span class=\"hljs-built_in\">print<\/span>(X1)\n<span class=\"hljs-built_in\">print<\/span>(X2)\n<span class=\"hljs-built_in\">print<\/span>(Y)\n<\/code><\/pre>\n<p>\u062f\u0631 \u0627\u0633\u06a9\u0631\u06cc\u067e\u062a \u0628\u0627\u0644\u0627\u060c \u0645\u0627 \u0633\u0647 \u0644\u06cc\u0633\u062a \u0627\u06cc\u062c\u0627\u062f \u0645\u06cc \u06a9\u0646\u06cc\u0645: <code>X1<\/code>\u060c <code>X2<\/code>\u060c \u0648 <code>Y<\/code>.  \u0647\u0631 \u0644\u06cc\u0633\u062a \u062f\u0627\u0631\u0627\u06cc 25 \u0639\u0646\u0635\u0631 \u0627\u0633\u062a\u060c \u0628\u0647 \u0627\u06cc\u0646 \u0645\u0639\u0646\u06cc \u06a9\u0647 \u062d\u062c\u0645 \u06a9\u0644 \u0646\u0645\u0648\u0646\u0647 25 \u0627\u0633\u062a. <code>Y<\/code> \u0634\u0627\u0645\u0644 \u062e\u0631\u0648\u062c\u06cc \u0645\u06cc \u0628\u0627\u0634\u062f. <code>X1<\/code>\u060c <code>X2<\/code>\u060c \u0648 <code>Y<\/code> \u0644\u06cc\u0633\u062a \u0647\u0627 \u062f\u0631 \u0632\u06cc\u0631 \u0686\u0627\u067e \u0634\u062f\u0647 \u0627\u0633\u062a:<\/p>\n<pre><code class=\"hljs\">(2, 4, 6, 8, 10, 12, 14, 16, 18, 20, 22, 24, 26, 28, 30, 32, 34, 36, 38, 40, 42, 44, 46, 48, 50)\n(3, 6, 9, 12, 15, 18, 21, 24, 27, 30, 33, 36, 39, 42, 45, 48, 51, 54, 57, 60, 63, 66, 69, 72, 75)\n(6, 24, 54, 96, 150, 216, 294, 384, 486, 600, 726, 864, 1014, 1176, 1350, 1536, 1734, 1944, 2166, 2400, 2646, 2904, 3174, 3456, 3750)\n<\/code><\/pre>\n<p>\u0647\u0631 \u0639\u0646\u0635\u0631 \u062f\u0631 \u0644\u06cc\u0633\u062a \u062e\u0631\u0648\u062c\u06cc \u0627\u0633\u0627\u0633\u0627\u064b \u062d\u0627\u0635\u0644 \u0636\u0631\u0628 \u0639\u0646\u0627\u0635\u0631 \u0645\u0631\u0628\u0648\u0637\u0647 \u062f\u0631 \u0641\u0647\u0631\u0633\u062a \u0627\u0633\u062a <code>X1<\/code> \u0648 <code>X2<\/code> \u0644\u06cc\u0633\u062a \u0647\u0627  \u0628\u0647 \u0639\u0646\u0648\u0627\u0646 \u0645\u062b\u0627\u0644\u060c \u0639\u0646\u0635\u0631 \u062f\u0648\u0645 \u062f\u0631 \u0644\u06cc\u0633\u062a \u062e\u0631\u0648\u062c\u06cc 24 \u0627\u0633\u062a \u06a9\u0647 \u062d\u0627\u0635\u0644\u0636\u0631\u0628 \u0639\u0646\u0635\u0631 \u062f\u0648\u0645 \u062f\u0631 \u0644\u06cc\u0633\u062a \u0627\u0633\u062a. <code>X1<\/code> \u06cc\u0639\u0646\u06cc 4 \u0648 \u0639\u0646\u0635\u0631 \u062f\u0648\u0645 \u062f\u0631 \u0644\u06cc\u0633\u062a <code>X2<\/code> \u06cc\u0639\u0646\u06cc 6.<\/p>\n<p>\u0648\u0631\u0648\u062f\u06cc \u0627\u0632 \u062a\u0631\u06a9\u06cc\u0628\u06cc \u0627\u0632 <code>X1<\/code> \u0648 <code>X2<\/code> \u0644\u06cc\u0633\u062a \u0647\u0627\u060c \u06a9\u0647 \u062f\u0631 \u0622\u0646 \u0647\u0631 \u0644\u06cc\u0633\u062a \u0628\u0647 \u0639\u0646\u0648\u0627\u0646 \u06cc\u06a9 \u0633\u062a\u0648\u0646 \u0646\u0634\u0627\u0646 \u062f\u0627\u062f\u0647 \u0645\u06cc \u0634\u0648\u062f.  \u0627\u0633\u06a9\u0631\u06cc\u067e\u062a \u0632\u06cc\u0631 \u0648\u0631\u0648\u062f\u06cc \u0646\u0647\u0627\u06cc\u06cc \u0631\u0627 \u0627\u06cc\u062c\u0627\u062f \u0645\u06cc \u06a9\u0646\u062f:<\/p>\n<pre><code class=\"hljs\">X = np.column_stack((X1, X2))\n<span class=\"hljs-built_in\">print<\/span>(X)\n<\/code><\/pre>\n<p>\u062f\u0631 \u0627\u06cc\u0646\u062c\u0627 \u062e\u0631\u0648\u062c\u06cc \u0627\u0633\u062a:<\/p>\n<pre><code class=\"hljs\">(( 2  3)\n ( 4  6)\n ( 6  9)\n ( 8 12)\n (10 15)\n (12 18)\n (14 21)\n (16 24)\n (18 27)\n (20 30)\n (22 33)\n (24 36)\n (26 39)\n (28 42)\n (30 45)\n (32 48)\n (34 51)\n (36 54)\n (38 57)\n (40 60)\n (42 63)\n (44 66)\n (46 69)\n (48 72)\n (50 75))\n<\/code><\/pre>\n<p>\u0627\u06cc\u0646\u062c\u0627 <code>X<\/code> \u0645\u062a\u063a\u06cc\u0631 \u0634\u0627\u0645\u0644 \u0645\u062c\u0645\u0648\u0639\u0647 \u0648\u06cc\u0698\u06af\u06cc \u0647\u0627\u06cc \u0646\u0647\u0627\u06cc\u06cc \u0645\u0627 \u0627\u0633\u062a.  \u0645\u06cc \u062a\u0648\u0627\u0646\u06cc\u062f \u0628\u0628\u06cc\u0646\u06cc\u062f \u06a9\u0647 \u0634\u0627\u0645\u0644 \u062f\u0648 \u0633\u062a\u0648\u0646 \u06cc\u0639\u0646\u06cc \u062f\u0648 \u0648\u06cc\u0698\u06af\u06cc \u062f\u0631 \u0647\u0631 \u0648\u0631\u0648\u062f\u06cc \u0627\u0633\u062a.  \u0647\u0645\u0627\u0646\u0637\u0648\u0631 \u06a9\u0647 \u0642\u0628\u0644\u0627\u064b \u0635\u062d\u0628\u062a \u06a9\u0631\u062f\u06cc\u0645\u060c \u0628\u0627\u06cc\u062f \u0648\u0631\u0648\u062f\u06cc \u0631\u0627 \u0628\u0647 \u06cc\u06a9 \u0634\u06a9\u0644 3 \u0628\u0639\u062f\u06cc \u062a\u0628\u062f\u06cc\u0644 \u06a9\u0646\u06cc\u0645.  \u0648\u0631\u0648\u062f\u06cc \u0645\u0627 \u062f\u0627\u0631\u0627\u06cc 25 \u0646\u0645\u0648\u0646\u0647 \u0627\u0633\u062a \u06a9\u0647 \u0647\u0631 \u0646\u0645\u0648\u0646\u0647 \u0627\u0632 1 \u0645\u0631\u062d\u0644\u0647 \u0632\u0645\u0627\u0646\u06cc \u0648 \u0647\u0631 \u0645\u0631\u062d\u0644\u0647 \u0632\u0645\u0627\u0646\u06cc \u0634\u0627\u0645\u0644 2 \u0648\u06cc\u0698\u06af\u06cc \u0627\u0633\u062a.  \u0627\u0633\u06a9\u0631\u06cc\u067e\u062a \u0632\u06cc\u0631 \u0648\u0631\u0648\u062f\u06cc \u0631\u0627 \u062a\u063a\u06cc\u06cc\u0631 \u0634\u06a9\u0644 \u0645\u06cc \u062f\u0647\u062f.<\/p>\n<pre><code class=\"hljs\">X = array(X).reshape(<span class=\"hljs-number\">25<\/span>, <span class=\"hljs-number\">1<\/span>, <span class=\"hljs-number\">2<\/span>)\n<\/code><\/pre>\n<h4 id=\"solutionviasimplelstm\">\u0631\u0627\u0647 \u062d\u0644 \u0627\u0632 \u0637\u0631\u06cc\u0642 LSTM \u0633\u0627\u062f\u0647<\/h4>\n<p>\u0645\u0627 \u0627\u06a9\u0646\u0648\u0646 \u0622\u0645\u0627\u062f\u0647 \u0622\u0645\u0648\u0632\u0634 \u0645\u062f\u0644 \u0647\u0627\u06cc LSTM \u062e\u0648\u062f \u0647\u0633\u062a\u06cc\u0645.  \u0628\u06cc\u0627\u06cc\u06cc\u062f \u0627\u0628\u062a\u062f\u0627 \u06cc\u06a9 \u0645\u062f\u0644 \u0644\u0627\u06cc\u0647 \u062a\u06a9 LSTM \u0631\u0627 \u0647\u0645\u0627\u0646\u0637\u0648\u0631 \u06a9\u0647 \u062f\u0631 \u0628\u062e\u0634 \u0642\u0628\u0644 \u0627\u0646\u062c\u0627\u0645 \u062f\u0627\u062f\u06cc\u0645 \u062a\u0648\u0633\u0639\u0647 \u062f\u0647\u06cc\u0645:<\/p>\n<pre><code class=\"hljs\">model = Sequential()\nmodel.add(LSTM(<span class=\"hljs-number\">80<\/span>, activation=<span class=\"hljs-string\">'relu'<\/span>, input_shape=(<span class=\"hljs-number\">1<\/span>, <span class=\"hljs-number\">2<\/span>)))\nmodel.add(Dense(<span class=\"hljs-number\">10<\/span>, activation=<span class=\"hljs-string\">'relu'<\/span>))\nmodel.add(Dense(<span class=\"hljs-number\">1<\/span>))\nmodel.<span class=\"hljs-built_in\">compile<\/span>(optimizer=<span class=\"hljs-string\">'adam'<\/span>, loss=<span class=\"hljs-string\">'mse'<\/span>)\n<span class=\"hljs-built_in\">print<\/span>(model.summary())\n<\/code><\/pre>\n<p>\u062f\u0631 \u0627\u06cc\u0646\u062c\u0627 \u0644\u0627\u06cc\u0647 LSTM \u0645\u0627 \u0634\u0627\u0645\u0644 80 \u0646\u0648\u0631\u0648\u0646 \u0627\u0633\u062a.  \u0645\u0627 \u062f\u0648 \u0644\u0627\u06cc\u0647 \u0645\u062a\u0631\u0627\u06a9\u0645 \u062f\u0627\u0631\u06cc\u0645 \u06a9\u0647 \u0644\u0627\u06cc\u0647 \u0627\u0648\u0644 \u0634\u0627\u0645\u0644 10 \u0646\u0648\u0631\u0648\u0646 \u0648 \u0644\u0627\u06cc\u0647 \u0645\u062a\u0631\u0627\u06a9\u0645 \u062f\u0648\u0645 \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 \u0634\u0627\u0645\u0644 1 \u0646\u0648\u0631\u0648\u0646 \u0627\u0633\u062a.  \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 #\n=================================================================\nlstm_38 (LSTM)               (None, 80)                26560\n_________________________________________________________________\ndense_29 (Dense)             (None, 10)                810\n_________________________________________________________________\ndense_30 (Dense)             (None, 1)                 11\n=================================================================\nTotal params: 27,381\nTrainable params: 27,381\nNon-trainable params: 0\n_________________________________________________________________\nNone\n<\/code><\/pre>\n<p>\u0627\u0633\u06a9\u0631\u06cc\u067e\u062a \u0632\u06cc\u0631 \u0645\u062f\u0644 \u0631\u0627 \u0622\u0645\u0648\u0632\u0634 \u0645\u06cc \u062f\u0647\u062f:<\/p>\n<pre><code class=\"hljs\">model.fit(X, Y, epochs=<span class=\"hljs-number\">2000<\/span>, validation_split=<span class=\"hljs-number\">0.2<\/span>, batch_size=<span class=\"hljs-number\">5<\/span>)\n<\/code><\/pre>\n<p>\u0628\u06cc\u0627\u06cc\u06cc\u062f \u0645\u062f\u0644 \u0622\u0645\u0648\u0632\u0634 \u062f\u06cc\u062f\u0647 \u062e\u0648\u062f \u0631\u0627 \u0622\u0632\u0645\u0627\u06cc\u0634 \u06a9\u0646\u06cc\u0645 \u0631\u0648\u06cc \u06cc\u06a9 \u0646\u0642\u0637\u0647 \u062f\u0627\u062f\u0647 \u062c\u062f\u06cc\u062f  \u0646\u0642\u0637\u0647 \u062f\u0627\u062f\u0647 \u0645\u0627 \u062f\u0648 \u0648\u06cc\u0698\u06af\u06cc \u062e\u0648\u0627\u0647\u062f \u062f\u0627\u0634\u062a\u060c \u06cc\u0639\u0646\u06cc (55\u060c80) \u062e\u0631\u0648\u062c\u06cc \u0648\u0627\u0642\u0639\u06cc \u0628\u0627\u06cc\u062f 55 x 80 = 4400 \u0628\u0627\u0634\u062f. \u0628\u06cc\u0627\u06cc\u06cc\u062f \u0628\u0628\u06cc\u0646\u06cc\u0645 \u0627\u0644\u06af\u0648\u0631\u06cc\u062a\u0645 \u0645\u0627 \u0686\u0647 \u0686\u06cc\u0632\u06cc \u0631\u0627 \u067e\u06cc\u0634 \u0628\u06cc\u0646\u06cc \u0645\u06cc \u06a9\u0646\u062f.  \u0627\u0633\u06a9\u0631\u06cc\u067e\u062a \u0632\u06cc\u0631 \u0631\u0627 \u0627\u062c\u0631\u0627 \u06a9\u0646\u06cc\u062f:<\/p>\n<pre><code class=\"hljs\">test_input = array((<span class=\"hljs-number\">55<\/span>,<span class=\"hljs-number\">80<\/span>))\ntest_input = test_input.reshape((<span class=\"hljs-number\">1<\/span>, <span class=\"hljs-number\">1<\/span>, <span class=\"hljs-number\">2<\/span>))\ntest_output = model.predict(test_input, verbose=<span class=\"hljs-number\">0<\/span>)\n<span class=\"hljs-built_in\">print<\/span>(test_output)\n<\/code><\/pre>\n<p>\u0645\u0646 \u062f\u0631 \u062e\u0631\u0648\u062c\u06cc 3263.44 \u06af\u0631\u0641\u062a\u0645 \u06a9\u0647 \u0628\u0627 \u062e\u0631\u0648\u062c\u06cc \u0648\u0627\u0642\u0639\u06cc \u0641\u0627\u0635\u0644\u0647 \u0632\u06cc\u0627\u062f\u06cc \u062f\u0627\u0631\u062f.<\/p>\n<h4 id=\"solutionviastackedlstm\">\u0631\u0627\u0647 \u062d\u0644 \u0627\u0632 \u0637\u0631\u06cc\u0642 Stacked LSTM<\/h4>\n<p>\u0628\u06cc\u0627\u06cc\u06cc\u062f \u0627\u06a9\u0646\u0648\u0646 \u06cc\u06a9 LSTM \u067e\u06cc\u0686\u06cc\u062f\u0647 \u062a\u0631 \u0628\u0627 \u0686\u0646\u062f\u06cc\u0646 LSTM \u0648 \u0644\u0627\u06cc\u0647 \u0647\u0627\u06cc \u0645\u062a\u0631\u0627\u06a9\u0645 \u0627\u06cc\u062c\u0627\u062f \u06a9\u0646\u06cc\u0645 \u0648 \u0628\u0628\u06cc\u0646\u06cc\u0645 \u06a9\u0647 \u0622\u06cc\u0627 \u0645\u06cc \u062a\u0648\u0627\u0646\u06cc\u0645 \u067e\u0627\u0633\u062e \u062e\u0648\u062f \u0631\u0627 \u0628\u0647\u0628\u0648\u062f \u062f\u0647\u06cc\u0645:<\/p>\n<pre><code class=\"hljs\">model = Sequential()\nmodel.add(LSTM(<span class=\"hljs-number\">200<\/span>, activation=<span class=\"hljs-string\">'relu'<\/span>, return_sequences=<span class=\"hljs-literal\">True<\/span>, input_shape=(<span class=\"hljs-number\">1<\/span>, <span class=\"hljs-number\">2<\/span>)))\nmodel.add(LSTM(<span class=\"hljs-number\">100<\/span>, activation=<span class=\"hljs-string\">'relu'<\/span>, return_sequences=<span class=\"hljs-literal\">True<\/span>))\nmodel.add(LSTM(<span class=\"hljs-number\">50<\/span>, activation=<span class=\"hljs-string\">'relu'<\/span>, return_sequences=<span class=\"hljs-literal\">True<\/span>))\nmodel.add(LSTM(<span class=\"hljs-number\">25<\/span>, activation=<span class=\"hljs-string\">'relu'<\/span>))\nmodel.add(Dense(<span class=\"hljs-number\">20<\/span>, activation=<span class=\"hljs-string\">'relu'<\/span>))\nmodel.add(Dense(<span class=\"hljs-number\">10<\/span>, activation=<span class=\"hljs-string\">'relu'<\/span>))\nmodel.add(Dense(<span class=\"hljs-number\">1<\/span>))\nmodel.<span class=\"hljs-built_in\">compile<\/span>(optimizer=<span class=\"hljs-string\">'adam'<\/span>, loss=<span class=\"hljs-string\">'mse'<\/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\">_________________________________________________________________\nLayer (type)                 Output Shape              Param #\n=================================================================\nlstm_53 (LSTM)               (None, 1, 200)            162400\n_________________________________________________________________\nlstm_54 (LSTM)               (None, 1, 100)            120400\n_________________________________________________________________\nlstm_55 (LSTM)               (None, 1, 50)             30200\n_________________________________________________________________\nlstm_56 (LSTM)               (None, 25)                7600\n_________________________________________________________________\ndense_43 (Dense)             (None, 20)                520\n_________________________________________________________________\ndense_44 (Dense)             (None, 10)                210\n_________________________________________________________________\ndense_45 (Dense)             (None, 1)                 11\n=================================================================\nTotal params: 321,341\nTrainable params: 321,341\nNon-trainable params: 0\n<\/code><\/pre>\n<p>\u0645\u0631\u062d\u0644\u0647 \u0628\u0639\u062f\u06cc \u0622\u0645\u0648\u0632\u0634 \u0645\u062f\u0644 \u0648 \u062a\u0633\u062a \u0622\u0646 \u0627\u0633\u062a \u0631\u0648\u06cc \u0646\u0642\u0637\u0647 \u062f\u0627\u062f\u0647 \u0622\u0632\u0645\u0648\u0646 \u06cc\u0639\u0646\u06cc (55\u060c80).<\/p>\n<p>\u0628\u0631\u0627\u06cc \u0628\u0647\u0628\u0648\u062f \u062f\u0642\u062a\u060c \u0627\u0646\u062f\u0627\u0632\u0647 \u062f\u0633\u062a\u0647 \u0631\u0627 \u06a9\u0627\u0647\u0634 \u0645\u06cc\u200c\u062f\u0647\u06cc\u0645 \u0648 \u0627\u0632 \u0622\u0646\u062c\u0627\u06cc\u06cc \u06a9\u0647 \u0645\u062f\u0644 \u0645\u0627 \u0627\u06a9\u0646\u0648\u0646 \u067e\u06cc\u0686\u06cc\u062f\u0647\u200c\u062a\u0631 \u0627\u0633\u062a\u060c \u0645\u06cc\u200c\u062a\u0648\u0627\u0646\u06cc\u0645 \u062a\u0639\u062f\u0627\u062f \u062f\u0648\u0631\u0647\u200c\u0647\u0627 \u0631\u0627 \u0646\u06cc\u0632 \u06a9\u0627\u0647\u0634 \u062f\u0647\u06cc\u0645.  \u0627\u0633\u06a9\u0631\u06cc\u067e\u062a \u0632\u06cc\u0631 \u0645\u062f\u0644 LSTM \u0631\u0627 \u0622\u0645\u0648\u0632\u0634 \u0645\u06cc \u062f\u0647\u062f \u0648 \u067e\u06cc\u0634 \u0628\u06cc\u0646\u06cc \u0645\u06cc \u06a9\u0646\u062f \u0631\u0648\u06cc \u0646\u0642\u0637\u0647 \u062f\u0627\u062f\u0647 \u062a\u0633\u062a<\/p>\n<pre><code class=\"hljs\">history = model.fit(X, Y, epochs=<span class=\"hljs-number\">1000<\/span>, validation_split=<span class=\"hljs-number\">0.1<\/span>, verbose=<span class=\"hljs-number\">1<\/span>, batch_size=<span class=\"hljs-number\">3<\/span>)\n\ntest_output = model.predict(test_input, verbose=<span class=\"hljs-number\">0<\/span>)\n<span class=\"hljs-built_in\">print<\/span>(test_output)\n<\/code><\/pre>\n<p>\u062f\u0631 \u062e\u0631\u0648\u062c\u06cc\u060c \u0645\u0642\u062f\u0627\u0631 3705.33 \u0631\u0627 \u062f\u0631\u06cc\u0627\u0641\u062a \u06a9\u0631\u062f\u0645 \u06a9\u0647 \u0647\u0646\u0648\u0632 \u06a9\u0645\u062a\u0631 \u0627\u0632 4400 \u0627\u0633\u062a\u060c \u0627\u0645\u0627 \u0628\u0633\u06cc\u0627\u0631 \u0628\u0647\u062a\u0631 \u0627\u0632 \u0645\u0642\u062f\u0627\u0631 \u0642\u0628\u0644\u06cc 3263.44 \u0628\u0627 \u0627\u0633\u062a\u0641\u0627\u062f\u0647 \u0627\u0632 \u06cc\u06a9 \u0644\u0627\u06cc\u0647 LSTM \u0645\u0646\u0641\u0631\u062f \u0627\u0633\u062a.  \u0645\u06cc\u200c\u062a\u0648\u0627\u0646\u06cc\u062f \u0628\u0627 \u062a\u0631\u06a9\u06cc\u0628\u200c\u0647\u0627\u06cc \u0645\u062e\u062a\u0644\u0641 \u0644\u0627\u06cc\u0647\u200c\u0647\u0627\u06cc LSTM\u060c \u0644\u0627\u06cc\u0647\u200c\u0647\u0627\u06cc \u0645\u062a\u0631\u0627\u06a9\u0645\u060c \u0627\u0646\u062f\u0627\u0632\u0647 \u062f\u0633\u062a\u0647\u200c\u0627\u06cc \u0648 \u062a\u0639\u062f\u0627\u062f \u062f\u0648\u0631\u0647\u200c\u0647\u0627 \u0628\u0627\u0632\u06cc \u06a9\u0646\u06cc\u062f \u062a\u0627 \u0628\u0628\u06cc\u0646\u06cc\u062f \u0622\u06cc\u0627 \u0646\u062a\u0627\u06cc\u062c \u0628\u0647\u062a\u0631\u06cc \u0645\u06cc\u200c\u06af\u06cc\u0631\u06cc\u062f \u06cc\u0627 \u062e\u06cc\u0631.<\/p>\n<h2 id=\"manytoonesequenceproblems\"><span class=\"ez-toc-section\" id=\"%d9%85%d8%b4%da%a9%d9%84%d8%a7%d8%aa_%d8%aa%d9%88%d8%a7%d9%84%db%8c_%da%86%d9%86%d8%af_%d8%a8%d9%87_%db%8c%da%a9\"><\/span>\u0645\u0634\u06a9\u0644\u0627\u062a \u062a\u0648\u0627\u0644\u06cc \u0686\u0646\u062f \u0628\u0647 \u06cc\u06a9<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<p>\u062f\u0631 \u0628\u062e\u0634 \u0647\u0627\u06cc \u0642\u0628\u0644\u06cc \u0631\u0648\u0634 \u062d\u0644 \u0645\u0633\u0627\u0626\u0644 \u06cc\u06a9 \u0628\u0647 \u06cc\u06a9 \u062a\u0648\u0627\u0644\u06cc \u0628\u0627 LSTM \u0631\u0627 \u062f\u06cc\u062f\u06cc\u0645.  \u062f\u0631 \u06cc\u06a9 \u0645\u0633\u0626\u0644\u0647 \u062a\u0648\u0627\u0644\u06cc \u06cc\u06a9 \u0628\u0647 \u06cc\u06a9\u060c \u0647\u0631 \u0646\u0645\u0648\u0646\u0647 \u0627\u0632 \u06cc\u06a9 \u0645\u0631\u062d\u0644\u0647 \u0632\u0645\u0627\u0646\u06cc \u0648\u0627\u062d\u062f \u0627\u0632 \u06cc\u06a9 \u06cc\u0627 \u0686\u0646\u062f \u0648\u06cc\u0698\u06af\u06cc \u062a\u0634\u06a9\u06cc\u0644 \u0634\u062f\u0647 \u0627\u0633\u062a.  \u062f\u0627\u062f\u0647 \u0647\u0627\u06cc\u06cc \u0628\u0627 \u06cc\u06a9 \u0645\u0631\u062d\u0644\u0647 \u0632\u0645\u0627\u0646\u06cc \u0648\u0627\u062d\u062f \u0631\u0627 \u0646\u0645\u06cc \u062a\u0648\u0627\u0646 \u0628\u0647 \u0645\u0639\u0646\u0627\u06cc \u0648\u0627\u0642\u0639\u06cc \u062f\u0627\u062f\u0647 \u0647\u0627\u06cc \u062a\u0648\u0627\u0644\u06cc \u062f\u0631 \u0646\u0638\u0631 \u06af\u0631\u0641\u062a.  \u062b\u0627\u0628\u062a \u0634\u062f\u0647 \u0627\u0633\u062a \u06a9\u0647 \u0634\u0628\u06a9\u0647 \u0647\u0627\u06cc \u0639\u0635\u0628\u06cc \u0645\u062a\u0635\u0644 \u0645\u062a\u0631\u0627\u06a9\u0645 \u0628\u0627 \u062f\u0627\u062f\u0647 \u0647\u0627\u06cc \u062a\u06a9 \u0645\u0631\u062d\u0644\u0647 \u0627\u06cc \u0639\u0645\u0644\u06a9\u0631\u062f \u0628\u0647\u062a\u0631\u06cc \u062f\u0627\u0631\u0646\u062f.<\/p>\n<p>\u062f\u0627\u062f\u0647 \u0647\u0627\u06cc \u062f\u0646\u0628\u0627\u0644\u0647 \u0648\u0627\u0642\u0639\u06cc \u0634\u0627\u0645\u0644 \u0686\u0646\u062f\u06cc\u0646 \u0645\u0631\u062d\u0644\u0647 \u0632\u0645\u0627\u0646\u06cc \u0627\u0633\u062a\u060c \u0645\u0627\u0646\u0646\u062f \u0642\u06cc\u0645\u062a \u0647\u0627\u06cc \u0628\u0627\u0632\u0627\u0631 \u0633\u0647\u0627\u0645 \u062f\u0631 7 \u0631\u0648\u0632 \u06af\u0630\u0634\u062a\u0647\u060c \u06cc\u06a9 \u062c\u0645\u0644\u0647 \u062d\u0627\u0648\u06cc \u06a9\u0644\u0645\u0627\u062a \u0645\u062a\u0639\u062f\u062f \u0648 \u063a\u06cc\u0631\u0647. \u0631\u0648\u06cc.<\/p>\n<p>\u062f\u0631 \u0627\u06cc\u0646 \u0628\u062e\u0634 \u0631\u0648\u0634 \u062d\u0644 \u0645\u0633\u0627\u0626\u0644 \u062a\u0648\u0627\u0644\u06cc \u0686\u0646\u062f \u0628\u0647 \u06cc\u06a9 \u0631\u0627 \u062e\u0648\u0627\u0647\u06cc\u0645 \u062f\u06cc\u062f.  \u062f\u0631 \u0645\u0633\u0627\u0626\u0644 \u062a\u0648\u0627\u0644\u06cc \u0686\u0646\u062f \u0628\u0647 \u06cc\u06a9\u060c \u0647\u0631 \u0646\u0645\u0648\u0646\u0647 \u0648\u0631\u0648\u062f\u06cc \u0628\u06cc\u0634 \u0627\u0632 \u06cc\u06a9 \u0645\u0631\u062d\u0644\u0647 \u0632\u0645\u0627\u0646\u06cc \u062f\u0627\u0631\u062f\u060c \u0628\u0627 \u0627\u06cc\u0646 \u062d\u0627\u0644 \u062e\u0631\u0648\u062c\u06cc \u0627\u0632 \u06cc\u06a9 \u0639\u0646\u0635\u0631 \u062a\u0634\u06a9\u06cc\u0644 \u0634\u062f\u0647 \u0627\u0633\u062a.  \u0647\u0631 \u0645\u0631\u062d\u0644\u0647 \u0632\u0645\u0627\u0646\u06cc \u062f\u0631 \u0648\u0631\u0648\u062f\u06cc \u0645\u06cc \u062a\u0648\u0627\u0646\u062f \u06cc\u06a9 \u06cc\u0627 \u0686\u0646\u062f \u0648\u06cc\u0698\u06af\u06cc \u062f\u0627\u0634\u062a\u0647 \u0628\u0627\u0634\u062f.  \u0645\u0627 \u0628\u0627 \u0645\u0633\u0627\u0626\u0644 \u062a\u0648\u0627\u0644\u06cc \u0686\u0646\u062f \u0628\u0647 \u06cc\u06a9 \u06a9\u0647 \u062f\u0627\u0631\u0627\u06cc \u06cc\u06a9 \u0648\u06cc\u0698\u06af\u06cc \u0647\u0633\u062a\u0646\u062f \u0634\u0631\u0648\u0639 \u0645\u06cc \u06a9\u0646\u06cc\u0645\u060c \u0648 \u0633\u067e\u0633 \u062e\u0648\u0627\u0647\u06cc\u0645 \u062f\u06cc\u062f \u06a9\u0647 \u0686\u06af\u0648\u0646\u0647 \u0645\u06cc \u062a\u0648\u0627\u0646 \u0645\u0633\u0627\u0626\u0644 \u0686\u0646\u062f \u0628\u0647 \u06cc\u06a9 \u0631\u0627 \u062d\u0644 \u06a9\u0631\u062f \u06a9\u0647 \u062f\u0631 \u0622\u0646 \u06af\u0627\u0645 \u0647\u0627\u06cc \u0632\u0645\u0627\u0646\u06cc \u0648\u0631\u0648\u062f\u06cc \u062f\u0627\u0631\u0627\u06cc \u0686\u0646\u062f\u06cc\u0646 \u0648\u06cc\u0698\u06af\u06cc \u0647\u0633\u062a\u0646\u062f.<\/p>\n<h3 id=\"manytoonesequenceproblemswithasinglefeature\"><span class=\"ez-toc-section\" id=\"%d9%85%d8%b4%da%a9%d9%84%d8%a7%d8%aa_%d8%aa%d9%88%d8%a7%d9%84%db%8c_%da%86%d9%86%d8%af_%d8%a8%d9%87_%db%8c%da%a9_%d8%a8%d8%a7_%db%8c%da%a9_%d9%88%db%8c%da%98%da%af%db%8c_%d9%88%d8%a7%d8%ad%d8%af\"><\/span>\u0645\u0634\u06a9\u0644\u0627\u062a \u062a\u0648\u0627\u0644\u06cc \u0686\u0646\u062f \u0628\u0647 \u06cc\u06a9 \u0628\u0627 \u06cc\u06a9 \u0648\u06cc\u0698\u06af\u06cc \u0648\u0627\u062d\u062f<span class=\"ez-toc-section-end\"><\/span><\/h3>\n<p>\u0628\u06cc\u0627\u06cc\u06cc\u062f \u0627\u0628\u062a\u062f\u0627 \u0645\u062c\u0645\u0648\u0639\u0647 \u062f\u0627\u062f\u0647 \u0631\u0627 \u0627\u06cc\u062c\u0627\u062f \u06a9\u0646\u06cc\u0645.  \u0645\u062c\u0645\u0648\u0639\u0647 \u062f\u0627\u062f\u0647 \u0645\u0627 \u0634\u0627\u0645\u0644 15 \u0646\u0645\u0648\u0646\u0647 \u062e\u0648\u0627\u0647\u062f \u0628\u0648\u062f.  \u0647\u0631 \u0646\u0645\u0648\u0646\u0647 \u062f\u0627\u0631\u0627\u06cc 3 \u0645\u0631\u062d\u0644\u0647 \u0632\u0645\u0627\u0646\u06cc \u062e\u0648\u0627\u0647\u062f \u0628\u0648\u062f \u06a9\u0647 \u0647\u0631 \u0645\u0631\u062d\u0644\u0647 \u0632\u0645\u0627\u0646\u06cc \u0627\u0632 \u06cc\u06a9 \u0648\u06cc\u0698\u06af\u06cc \u0648\u0627\u062d\u062f \u06cc\u0639\u0646\u06cc \u06cc\u06a9 \u0639\u062f\u062f \u062a\u0634\u06a9\u06cc\u0644 \u0634\u062f\u0647 \u0627\u0633\u062a.  \u062e\u0631\u0648\u062c\u06cc \u0647\u0631 \u0646\u0645\u0648\u0646\u0647\u060c \u0645\u062c\u0645\u0648\u0639 \u0627\u0639\u062f\u0627\u062f \u062f\u0631 \u0647\u0631 \u06cc\u06a9 \u0627\u0632 \u0633\u0647 \u0645\u0631\u062d\u0644\u0647 \u0632\u0645\u0627\u0646\u06cc \u062e\u0648\u0627\u0647\u062f \u0628\u0648\u062f.  \u0628\u0647 \u0639\u0646\u0648\u0627\u0646 \u0645\u062b\u0627\u0644\u060c \u0627\u06af\u0631 \u0646\u0645\u0648\u0646\u0647 \u0645\u0627 \u0634\u0627\u0645\u0644 \u062f\u0646\u0628\u0627\u0644\u0647 4\u060c5\u060c6 \u0628\u0627\u0634\u062f\u060c \u062e\u0631\u0648\u062c\u06cc 4 + 5 + 6 = 10 \u062e\u0648\u0627\u0647\u062f \u0628\u0648\u062f.<\/p>\n<h4 id=\"creatingthedataset\">\u0627\u06cc\u062c\u0627\u062f \u0645\u062c\u0645\u0648\u0639\u0647 \u062f\u0627\u062f\u0647<\/h4>\n<p>\u0628\u06cc\u0627\u06cc\u06cc\u062f \u0627\u0628\u062a\u062f\u0627 \u0644\u06cc\u0633\u062a\u06cc \u0627\u0632 \u0627\u0639\u062f\u0627\u062f \u0635\u062d\u06cc\u062d \u0627\u0632 1 \u062a\u0627 45 \u0627\u06cc\u062c\u0627\u062f \u06a9\u0646\u06cc\u0645. \u0627\u0632 \u0622\u0646\u062c\u0627\u06cc\u06cc \u06a9\u0647 \u0645\u0627 15 \u0646\u0645\u0648\u0646\u0647 \u062f\u0631 \u0645\u062c\u0645\u0648\u0639\u0647 \u062f\u0627\u062f\u0647 \u062e\u0648\u062f \u0645\u06cc \u062e\u0648\u0627\u0647\u06cc\u0645\u060c \u200b\u200b\u0644\u06cc\u0633\u062a\u06cc \u0627\u0632 \u0627\u0639\u062f\u0627\u062f \u0635\u062d\u06cc\u062d \u062d\u0627\u0648\u06cc 45 \u0639\u062f\u062f \u0635\u062d\u06cc\u062d \u0627\u0648\u0644 \u0631\u0627 \u062a\u063a\u06cc\u06cc\u0631 \u0634\u06a9\u0644 \u0645\u06cc \u062f\u0647\u06cc\u0645.<\/p>\n<pre><code class=\"hljs\">X = np.array((x+<span class=\"hljs-number\">1<\/span> <span class=\"hljs-keyword\">for<\/span> x <span class=\"hljs-keyword\">in<\/span> <span class=\"hljs-built_in\">range<\/span>(<span class=\"hljs-number\">45<\/span>)))\n<span class=\"hljs-built_in\">print<\/span>(X)\n<\/code><\/pre>\n<p>\u062f\u0631 \u062e\u0631\u0648\u062c\u06cc \u0628\u0627\u06cc\u062f 45 \u0639\u062f\u062f \u0635\u062d\u06cc\u062d \u0627\u0648\u0644 \u0631\u0627 \u0628\u0628\u06cc\u0646\u06cc\u062f:<\/p>\n<pre><code class=\"hljs\">( 1  2  3  4  5  6  7  8  9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24\n 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45)\n<\/code><\/pre>\n<p>\u0628\u0627 \u0627\u0633\u062a\u0641\u0627\u062f\u0647 \u0627\u0632 \u062a\u0627\u0628\u0639 \u0632\u06cc\u0631 \u0645\u06cc\u200c\u062a\u0648\u0627\u0646\u06cc\u0645 \u0622\u0646 \u0631\u0627 \u0628\u0647 \u062a\u0639\u062f\u0627\u062f\u06cc \u0646\u0645\u0648\u0646\u0647\u060c \u0645\u0631\u0627\u062d\u0644 \u0632\u0645\u0627\u0646\u06cc \u0648 \u0648\u06cc\u0698\u06af\u06cc\u200c\u0647\u0627 \u062a\u063a\u06cc\u06cc\u0631 \u0634\u06a9\u0644 \u062f\u0647\u06cc\u0645:<\/p>\n<pre><code class=\"hljs\">X = X.reshape(<span class=\"hljs-number\">15<\/span>,<span class=\"hljs-number\">3<\/span>,<span class=\"hljs-number\">1<\/span>)\n<span class=\"hljs-built_in\">print<\/span>(X)\n<\/code><\/pre>\n<p>\u0627\u0633\u06a9\u0631\u06cc\u067e\u062a \u0628\u0627\u0644\u0627 \u0644\u06cc\u0633\u062a \u0631\u0627 \u062a\u0628\u062f\u06cc\u0644 \u0645\u06cc \u06a9\u0646\u062f <code>X<\/code> \u0628\u0647 \u0634\u06a9\u0644 3 \u0628\u0639\u062f\u06cc \u0628\u0627 15 \u0646\u0645\u0648\u0646\u0647\u060c 3 \u0645\u0631\u062d\u0644\u0647 \u0632\u0645\u0627\u0646\u06cc \u0648 1 \u0648\u06cc\u0698\u06af\u06cc.  \u0627\u0633\u06a9\u0631\u06cc\u067e\u062a \u0628\u0627\u0644\u0627 \u0647\u0645\u0686\u0646\u06cc\u0646 \u062f\u0627\u062f\u0647 \u0647\u0627\u06cc \u062a\u063a\u06cc\u06cc\u0631 \u0634\u06a9\u0644 \u062f\u0627\u062f\u0647 \u0634\u062f\u0647 \u0631\u0627 \u0686\u0627\u067e \u0645\u06cc \u06a9\u0646\u062f.<\/p>\n<pre><code class=\"hljs\">((( 1)\n  ( 2)\n  ( 3))\n\n (( 4)\n  ( 5)\n  ( 6))\n\n (( 7)\n  ( 8)\n  ( 9))\n\n ((10)\n  (11)\n  (12))\n\n ((13)\n  (14)\n  (15))\n\n ((16)\n  (17)\n  (18))\n\n ((19)\n  (20)\n  (21))\n\n ((22)\n  (23)\n  (24))\n\n ((25)\n  (26)\n  (27))\n\n ((28)\n  (29)\n  (30))\n\n ((31)\n  (32)\n  (33))\n\n ((34)\n  (35)\n  (36))\n\n ((37)\n  (38)\n  (39))\n\n ((40)\n  (41)\n  (42))\n\n ((43)\n  (44)\n  (45)))\n<\/code><\/pre>\n<p>\u0645\u0627 \u062f\u0627\u062f\u0647 \u0647\u0627\u06cc \u0648\u0631\u0648\u062f\u06cc \u062e\u0648\u062f \u0631\u0627 \u0628\u0647 \u0641\u0631\u0645\u062a \u0645\u0646\u0627\u0633\u0628 \u062a\u0628\u062f\u06cc\u0644 \u06a9\u0631\u062f\u0647 \u0627\u06cc\u0645\u060c \u0627\u06a9\u0646\u0648\u0646 \u0628\u0631\u062f\u0627\u0631 \u062e\u0631\u0648\u062c\u06cc \u062e\u0648\u062f \u0631\u0627 \u0627\u06cc\u062c\u0627\u062f \u0645\u06cc \u06a9\u0646\u06cc\u0645.  \u0647\u0645\u0627\u0646\u0637\u0648\u0631 \u06a9\u0647 \u0642\u0628\u0644\u0627\u064b \u06af\u0641\u062a\u0645\u060c \u0647\u0631 \u0639\u0646\u0635\u0631 \u062f\u0631 \u062e\u0631\u0648\u062c\u06cc \u0628\u0631\u0627\u0628\u0631 \u0628\u0627 \u0645\u062c\u0645\u0648\u0639 \u0645\u0642\u0627\u062f\u06cc\u0631 \u062f\u0631 \u0645\u0631\u0627\u062d\u0644 \u0632\u0645\u0627\u0646\u06cc \u062f\u0631 \u0646\u0645\u0648\u0646\u0647 \u0648\u0631\u0648\u062f\u06cc \u0645\u0631\u0628\u0648\u0637\u0647 \u062e\u0648\u0627\u0647\u062f \u0628\u0648\u062f.  \u0627\u0633\u06a9\u0631\u06cc\u067e\u062a \u0632\u06cc\u0631 \u0628\u0631\u062f\u0627\u0631 \u062e\u0631\u0648\u062c\u06cc \u0631\u0627 \u0627\u06cc\u062c\u0627\u062f \u0645\u06cc \u06a9\u0646\u062f:<\/p>\n<pre><code class=\"hljs\">Y = <span class=\"hljs-built_in\">list<\/span>()\n<span class=\"hljs-keyword\">for<\/span> x <span class=\"hljs-keyword\">in<\/span> X:\n    Y.append(x.<span class=\"hljs-built_in\">sum<\/span>())\n\nY = np.array(Y)\n<span class=\"hljs-built_in\">print<\/span>(Y)\n<\/code><\/pre>\n<p>\u0622\u0631\u0627\u06cc\u0647 \u062e\u0631\u0648\u062c\u06cc <code>Y<\/code> \u0628\u0647 \u0646\u0638\u0631 \u0645\u06cc \u0631\u0633\u062f \u0627\u06cc\u0646 \u0627\u0633\u062a:<\/p>\n<pre><code class=\"hljs\">(  6  15  24  33  42  51  60  69  78  87  96 105 114 123 132)\n<\/code><\/pre>\n<h4 id=\"solutionviasimplelstm\">\u0631\u0627\u0647 \u062d\u0644 \u0627\u0632 \u0637\u0631\u06cc\u0642 LSTM \u0633\u0627\u062f\u0647<\/h4>\n<p>\u062d\u0627\u0644 \u0628\u06cc\u0627\u06cc\u06cc\u062f \u0645\u062f\u0644 \u062e\u0648\u062f \u0631\u0627 \u0628\u0627 \u06cc\u06a9 \u0644\u0627\u06cc\u0647 LSTM \u0627\u06cc\u062c\u0627\u062f \u06a9\u0646\u06cc\u0645.<\/p>\n<pre><code class=\"hljs\">model = Sequential()\nmodel.add(LSTM(<span class=\"hljs-number\">50<\/span>, activation=<span class=\"hljs-string\">'relu'<\/span>, input_shape=(<span class=\"hljs-number\">3<\/span>, <span class=\"hljs-number\">1<\/span>)))\nmodel.add(Dense(<span class=\"hljs-number\">1<\/span>))\nmodel.<span class=\"hljs-built_in\">compile<\/span>(optimizer=<span class=\"hljs-string\">'adam'<\/span>, loss=<span class=\"hljs-string\">'mse'<\/span>)\n<\/code><\/pre>\n<p>\u0627\u0633\u06a9\u0631\u06cc\u067e\u062a \u0632\u06cc\u0631 \u0645\u062f\u0644 \u0645\u0627 \u0631\u0627 \u0622\u0645\u0648\u0632\u0634 \u0645\u06cc \u062f\u0647\u062f:<\/p>\n<pre><code class=\"hljs\">history = model.fit(X, Y, epochs=<span class=\"hljs-number\">1000<\/span>, validation_split=<span class=\"hljs-number\">0.2<\/span>, verbose=<span class=\"hljs-number\">1<\/span>)\n<\/code><\/pre>\n<p>\u0647\u0646\u06af\u0627\u0645\u06cc \u06a9\u0647 \u0645\u062f\u0644 \u0622\u0645\u0648\u0632\u0634 \u062f\u0627\u062f\u0647 \u0634\u062f\u060c \u0645\u06cc \u062a\u0648\u0627\u0646\u06cc\u0645 \u0627\u0632 \u0622\u0646 \u0628\u0631\u0627\u06cc \u067e\u06cc\u0634 \u0628\u06cc\u0646\u06cc \u0627\u0633\u062a\u0641\u0627\u062f\u0647 \u06a9\u0646\u06cc\u0645 \u0631\u0648\u06cc \u0646\u0642\u0627\u0637 \u062f\u0627\u062f\u0647 \u062a\u0633\u062a  \u0628\u06cc\u0627\u06cc\u06cc\u062f \u062e\u0631\u0648\u062c\u06cc \u062f\u0646\u0628\u0627\u0644\u0647 \u0627\u0639\u062f\u0627\u062f 50\u060c51\u060c52 \u0631\u0627 \u067e\u06cc\u0634 \u0628\u06cc\u0646\u06cc \u06a9\u0646\u06cc\u0645.  \u062e\u0631\u0648\u062c\u06cc \u0648\u0627\u0642\u0639\u06cc \u0628\u0627\u06cc\u062f 50 + 51 + 52 = 153 \u0628\u0627\u0634\u062f. \u0627\u0633\u06a9\u0631\u06cc\u067e\u062a \u0632\u06cc\u0631 \u0646\u0642\u0627\u0637 \u062a\u0633\u062a \u0645\u0627 \u0631\u0627 \u0628\u0647 \u06cc\u06a9 \u0634\u06a9\u0644 3 \u0628\u0639\u062f\u06cc \u062a\u0628\u062f\u06cc\u0644 \u0645\u06cc \u06a9\u0646\u062f \u0648 \u0633\u067e\u0633 \u062e\u0631\u0648\u062c\u06cc \u0631\u0627 \u067e\u06cc\u0634 \u0628\u06cc\u0646\u06cc \u0645\u06cc \u06a9\u0646\u062f:<\/p>\n<pre><code class=\"hljs\">test_input = array((<span class=\"hljs-number\">50<\/span>,<span class=\"hljs-number\">51<\/span>,<span class=\"hljs-number\">52<\/span>))\ntest_input = test_input.reshape((<span class=\"hljs-number\">1<\/span>, <span class=\"hljs-number\">3<\/span>, <span class=\"hljs-number\">1<\/span>))\ntest_output = model.predict(test_input, verbose=<span class=\"hljs-number\">0<\/span>)\n<span class=\"hljs-built_in\">print<\/span>(test_output)\n<\/code><\/pre>\n<p>\u0645\u0646 145.96 \u062f\u0631 \u062e\u0631\u0648\u062c\u06cc \u06af\u0631\u0641\u062a\u0645 \u06a9\u0647 \u062d\u062f\u0648\u062f 7 \u0627\u0645\u062a\u06cc\u0627\u0632 \u06a9\u0645\u062a\u0631 \u0627\u0632 \u0645\u0642\u062f\u0627\u0631 \u0648\u0627\u0642\u0639\u06cc \u062e\u0631\u0648\u062c\u06cc 153 \u0627\u0633\u062a.<\/p>\n<h4 id=\"solutionviastackedlstm\">\u0631\u0627\u0647 \u062d\u0644 \u0627\u0632 \u0637\u0631\u06cc\u0642 Stacked LSTM<\/h4>\n<p>\u062d\u0627\u0644\u0627 \u0628\u06cc\u0627\u06cc\u06cc\u062f \u06cc\u06a9 \u0645\u062f\u0644 \u067e\u06cc\u0686\u06cc\u062f\u0647 LSTM \u0628\u0627 \u0686\u0646\u062f\u06cc\u0646 \u0644\u0627\u06cc\u0647 \u0627\u06cc\u062c\u0627\u062f \u06a9\u0646\u06cc\u0645 \u0648 \u0628\u0628\u06cc\u0646\u06cc\u0645 \u0622\u06cc\u0627 \u0645\u06cc\u200c\u062a\u0648\u0627\u0646\u06cc\u0645 \u0646\u062a\u0627\u06cc\u062c \u0628\u0647\u062a\u0631\u06cc \u0628\u06af\u06cc\u0631\u06cc\u0645.  \u0627\u0633\u06a9\u0631\u06cc\u067e\u062a \u0632\u06cc\u0631 \u0631\u0627 \u0628\u0631\u0627\u06cc \u0627\u06cc\u062c\u0627\u062f \u0648 \u0622\u0645\u0648\u0632\u0634 \u06cc\u06a9 \u0645\u062f\u0644 \u067e\u06cc\u0686\u06cc\u062f\u0647 \u0628\u0627 \u0686\u0646\u062f\u06cc\u0646 LSTM \u0648 \u0644\u0627\u06cc\u0647 \u0647\u0627\u06cc \u0645\u062a\u0631\u0627\u06a9\u0645 \u0627\u062c\u0631\u0627 \u06a9\u0646\u06cc\u062f:<\/p>\n<pre><code class=\"hljs\">model = Sequential()\nmodel.add(LSTM(<span class=\"hljs-number\">200<\/span>, activation=<span class=\"hljs-string\">'relu'<\/span>, return_sequences=<span class=\"hljs-literal\">True<\/span>, input_shape=(<span class=\"hljs-number\">3<\/span>, <span class=\"hljs-number\">1<\/span>)))\nmodel.add(LSTM(<span class=\"hljs-number\">100<\/span>, activation=<span class=\"hljs-string\">'relu'<\/span>, return_sequences=<span class=\"hljs-literal\">True<\/span>))\nmodel.add(LSTM(<span class=\"hljs-number\">50<\/span>, activation=<span class=\"hljs-string\">'relu'<\/span>, return_sequences=<span class=\"hljs-literal\">True<\/span>))\nmodel.add(LSTM(<span class=\"hljs-number\">25<\/span>, activation=<span class=\"hljs-string\">'relu'<\/span>))\nmodel.add(Dense(<span class=\"hljs-number\">20<\/span>, activation=<span class=\"hljs-string\">'relu'<\/span>))\nmodel.add(Dense(<span class=\"hljs-number\">10<\/span>, activation=<span class=\"hljs-string\">'relu'<\/span>))\nmodel.add(Dense(<span class=\"hljs-number\">1<\/span>))\nmodel.<span class=\"hljs-built_in\">compile<\/span>(optimizer=<span class=\"hljs-string\">'adam'<\/span>, loss=<span class=\"hljs-string\">'mse'<\/span>)\n\nhistory = model.fit(X, Y, epochs=<span class=\"hljs-number\">1000<\/span>, validation_split=<span class=\"hljs-number\">0.2<\/span>, verbose=<span class=\"hljs-number\">1<\/span>)\n<\/code><\/pre>\n<p>\u062d\u0627\u0644\u0627 \u0628\u06cc\u0627\u06cc\u06cc\u062f \u0645\u062f\u0644 \u062e\u0648\u062f \u0631\u0627 \u0622\u0632\u0645\u0627\u06cc\u0634 \u06a9\u0646\u06cc\u0645 \u0631\u0648\u06cc \u062f\u0646\u0628\u0627\u0644\u0647 \u0622\u0632\u0645\u0648\u0646 \u06cc\u0639\u0646\u06cc 50\u060c 51\u060c 52:<\/p>\n<pre><code class=\"hljs\">test_output = model.predict(test_input, verbose=<span class=\"hljs-number\">0<\/span>)\n<span class=\"hljs-built_in\">print<\/span>(test_output)\n<\/code><\/pre>\n<p>\u067e\u0627\u0633\u062e\u06cc \u06a9\u0647 \u0645\u0646 \u062f\u0631 \u0627\u06cc\u0646\u062c\u0627 \u06af\u0631\u0641\u062a\u0645 155.37 \u0627\u0633\u062a \u06a9\u0647 \u0628\u0647\u062a\u0631 \u0627\u0632 \u0646\u062a\u06cc\u062c\u0647 145.96 \u0627\u0633\u062a \u06a9\u0647 \u0642\u0628\u0644\u0627\u064b \u062f\u0631\u06cc\u0627\u0641\u062a \u06a9\u0631\u062f\u06cc\u0645.  \u062f\u0631 \u0627\u06cc\u0646 \u0635\u0648\u0631\u062a \u0627\u0632 153 \u0641\u0642\u0637 2 \u0627\u0645\u062a\u06cc\u0627\u0632 \u0627\u062e\u062a\u0644\u0627\u0641 \u062f\u0627\u0631\u06cc\u0645 \u06a9\u0647 \u067e\u0627\u0633\u062e \u0648\u0627\u0642\u0639\u06cc \u0627\u0633\u062a.<\/p>\n<h4 id=\"solutionviabidirectionallstm\">\u0631\u0627\u0647 \u062d\u0644 \u0627\u0632 \u0637\u0631\u06cc\u0642 \u062f\u0648 \u062c\u0647\u062a\u0647 LSTM<\/h4>\n<p><a rel=\"nofollow noopener\" target=\"_blank\" href=\"https:\/\/en.wikipedia.org\/wiki\/Bidirectional_recurrent_neural_networks\">LSTM \u062f\u0648 \u0637\u0631\u0641\u0647<\/a> \u0646\u0648\u0639\u06cc LSTM \u0627\u0633\u062a \u06a9\u0647 \u0627\u0632 \u062a\u0631\u062a\u06cc\u0628 \u0648\u0631\u0648\u062f\u06cc \u0627\u0632 \u0647\u0631 \u062f\u0648 \u062c\u0647\u062a \u062c\u0644\u0648 \u0648 \u0639\u0642\u0628 \u06cc\u0627\u062f \u0645\u06cc \u06af\u06cc\u0631\u062f.  \u062a\u0641\u0633\u06cc\u0631 \u062a\u0648\u0627\u0644\u06cc \u0646\u0647\u0627\u06cc\u06cc\u060c \u0627\u0644\u062d\u0627\u0642 \u0647\u0631 \u062f\u0648 \u067e\u0627\u0633 \u06cc\u0627\u062f\u06af\u06cc\u0631\u06cc \u0631\u0648 \u0628\u0647 \u062c\u0644\u0648 \u0648 \u0639\u0642\u0628 \u0627\u0633\u062a.  \u0628\u06cc\u0627\u06cc\u06cc\u062f \u0628\u0628\u06cc\u0646\u06cc\u0645 \u06a9\u0647 \u0622\u06cc\u0627 \u0645\u06cc \u062a\u0648\u0627\u0646\u06cc\u0645 \u0628\u0627 LSTM \u0647\u0627\u06cc \u062f\u0648 \u0637\u0631\u0641\u0647 \u0646\u062a\u0627\u06cc\u062c \u0628\u0647\u062a\u0631\u06cc \u0628\u06af\u06cc\u0631\u06cc\u0645 \u06cc\u0627 \u062e\u06cc\u0631.<\/p>\n<p>\u0627\u0633\u06a9\u0631\u06cc\u067e\u062a \u0632\u06cc\u0631 \u06cc\u06a9 \u0645\u062f\u0644 LSTM \u062f\u0648 \u0637\u0631\u0641\u0647 \u0628\u0627 \u06cc\u06a9 \u0644\u0627\u06cc\u0647 \u062f\u0648 \u0637\u0631\u0641\u0647 \u0648 \u06cc\u06a9 \u0644\u0627\u06cc\u0647 \u0645\u062a\u0631\u0627\u06a9\u0645 \u0627\u06cc\u062c\u0627\u062f \u0645\u06cc \u06a9\u0646\u062f \u06a9\u0647 \u0628\u0647 \u0639\u0646\u0648\u0627\u0646 \u062e\u0631\u0648\u062c\u06cc \u0645\u062f\u0644 \u0639\u0645\u0644 \u0645\u06cc \u06a9\u0646\u062f.<\/p>\n<pre><code class=\"hljs\"><span class=\"hljs-keyword\">from<\/span> keras.layers <span class=\"hljs-keyword\">import<\/span> Bidirectional\n\nmodel = Sequential()\nmodel.add(Bidirectional(LSTM(<span class=\"hljs-number\">50<\/span>, activation=<span class=\"hljs-string\">'relu'<\/span>), input_shape=(<span class=\"hljs-number\">3<\/span>, <span class=\"hljs-number\">1<\/span>)))\nmodel.add(Dense(<span class=\"hljs-number\">1<\/span>))\nmodel.<span class=\"hljs-built_in\">compile<\/span>(optimizer=<span class=\"hljs-string\">'adam'<\/span>, loss=<span class=\"hljs-string\">'mse'<\/span>)\n<\/code><\/pre>\n<p>\u0627\u0633\u06a9\u0631\u06cc\u067e\u062a \u0632\u06cc\u0631 \u0645\u062f\u0644 \u0631\u0627 \u0622\u0645\u0648\u0632\u0634 \u0645\u06cc \u062f\u0647\u062f \u0648 \u067e\u06cc\u0634 \u0628\u06cc\u0646\u06cc \u0645\u06cc \u06a9\u0646\u062f \u0631\u0648\u06cc \u062f\u0646\u0628\u0627\u0644\u0647 \u0622\u0632\u0645\u0648\u0646 \u06a9\u0647 50\u060c 51 \u0648 52 \u0627\u0633\u062a.<\/p>\n<pre><code class=\"hljs\">history = model.fit(X, Y, epochs=<span class=\"hljs-number\">1000<\/span>, validation_split=<span class=\"hljs-number\">0.2<\/span>, verbose=<span class=\"hljs-number\">1<\/span>)\ntest_output = model.predict(test_input, verbose=<span class=\"hljs-number\">0<\/span>)\n<span class=\"hljs-built_in\">print<\/span>(test_output)\n<\/code><\/pre>\n<p>\u0646\u062a\u06cc\u062c\u0647 \u0627\u06cc \u06a9\u0647 \u0645\u0646 \u06af\u0631\u0641\u062a\u0645 152.26 \u0627\u0633\u062a \u06a9\u0647 \u0641\u0642\u0637 \u06a9\u0633\u0631\u06cc \u0627\u0632 \u0646\u062a\u06cc\u062c\u0647 \u0648\u0627\u0642\u0639\u06cc \u06a9\u0645\u062a\u0631 \u0627\u0633\u062a.  \u0628\u0646\u0627\u0628\u0631\u0627\u06cc\u0646\u060c \u0645\u06cc\u200c\u062a\u0648\u0627\u0646 \u0646\u062a\u06cc\u062c\u0647 \u06af\u0631\u0641\u062a \u06a9\u0647 \u0628\u0631\u0627\u06cc \u0645\u062c\u0645\u0648\u0639\u0647 \u062f\u0627\u062f\u0647 \u0645\u0627\u060c LSTM \u062f\u0648\u0637\u0631\u0641\u0647 \u0628\u0627 \u062a\u06a9 \u0644\u0627\u06cc\u0647 \u0639\u0645\u0644\u06a9\u0631\u062f \u0628\u0647\u062a\u0631\u06cc \u0627\u0632 LSTM\u0647\u0627\u06cc \u062a\u06a9 \u0644\u0627\u06cc\u0647 \u0648 LSTM\u0647\u0627\u06cc \u06cc\u06a9 \u0637\u0631\u0641\u0647 \u0627\u0646\u0628\u0627\u0634\u062a\u0647 \u062f\u0627\u0631\u062f.<\/p>\n<h3 id=\"manytoonesequenceproblemswithmultiplefeatures\"><span class=\"ez-toc-section\" id=\"%d9%85%d8%b4%da%a9%d9%84%d8%a7%d8%aa_%d8%aa%d9%88%d8%a7%d9%84%db%8c_%da%86%d9%86%d8%af_%d8%a8%d9%87_%db%8c%da%a9_%d8%a8%d8%a7_%da%86%d9%86%d8%af%db%8c%d9%86_%d9%88%db%8c%da%98%da%af%db%8c\"><\/span>\u0645\u0634\u06a9\u0644\u0627\u062a \u062a\u0648\u0627\u0644\u06cc \u0686\u0646\u062f \u0628\u0647 \u06cc\u06a9 \u0628\u0627 \u0686\u0646\u062f\u06cc\u0646 \u0648\u06cc\u0698\u06af\u06cc<span class=\"ez-toc-section-end\"><\/span><\/h3>\n<p>\u062f\u0631 \u06cc\u06a9 \u0645\u0633\u0626\u0644\u0647 \u062a\u0648\u0627\u0644\u06cc \u0686\u0646\u062f \u0628\u0647 \u06cc\u06a9\u060c \u0648\u0631\u0648\u062f\u06cc \u062f\u0627\u0631\u06cc\u0645 \u06a9\u0647 \u062f\u0631 \u0622\u0646 \u0647\u0631 \u0645\u0631\u062d\u0644\u0647 \u0632\u0645\u0627\u0646\u06cc \u0627\u0632 \u0686\u0646\u062f\u06cc\u0646 \u0648\u06cc\u0698\u06af\u06cc \u062a\u0634\u06a9\u06cc\u0644 \u0634\u062f\u0647 \u0627\u0633\u062a.  \u062e\u0631\u0648\u062c\u06cc \u0645\u06cc \u062a\u0648\u0627\u0646\u062f \u06cc\u06a9 \u0645\u0642\u062f\u0627\u0631 \u0648\u0627\u062d\u062f \u06cc\u0627 \u0686\u0646\u062f \u0645\u0642\u062f\u0627\u0631 \u0628\u0627\u0634\u062f\u060c \u06cc\u06a9\u06cc \u062f\u0631 \u0647\u0631 \u0648\u06cc\u0698\u06af\u06cc \u062f\u0631 \u0645\u0631\u062d\u0644\u0647 \u0632\u0645\u0627\u0646\u06cc \u0648\u0631\u0648\u062f\u06cc.  \u062f\u0631 \u0627\u06cc\u0646 \u0628\u062e\u0634 \u0628\u0647 \u0647\u0631 \u062f\u0648 \u0645\u0648\u0631\u062f \u062e\u0648\u0627\u0647\u06cc\u0645 \u067e\u0631\u062f\u0627\u062e\u062a.<\/p>\n<h4 id=\"creatingthedataset\">\u0627\u06cc\u062c\u0627\u062f \u0645\u062c\u0645\u0648\u0639\u0647 \u062f\u0627\u062f\u0647<\/h4>\n<p>\u0645\u062c\u0645\u0648\u0639\u0647 \u062f\u0627\u062f\u0647 \u0645\u0627 \u0634\u0627\u0645\u0644 15 \u0646\u0645\u0648\u0646\u0647 \u062e\u0648\u0627\u0647\u062f \u0628\u0648\u062f.  \u0647\u0631 \u0646\u0645\u0648\u0646\u0647 \u0634\u0627\u0645\u0644 3 \u0645\u0631\u062d\u0644\u0647 \u0632\u0645\u0627\u0646\u06cc \u062e\u0648\u0627\u0647\u062f \u0628\u0648\u062f.  \u0647\u0631 time-step \u062f\u0648 \u0648\u06cc\u0698\u06af\u06cc \u062e\u0648\u0627\u0647\u062f \u062f\u0627\u0634\u062a.<\/p>\n<p>\u0628\u06cc\u0627\u06cc\u06cc\u062f \u062f\u0648 \u0644\u06cc\u0633\u062a \u0627\u06cc\u062c\u0627\u062f \u06a9\u0646\u06cc\u0645.  \u06cc\u06a9\u06cc \u0634\u0627\u0645\u0644 \u0645\u0636\u0631\u0628 \u0647\u0627\u06cc 3 \u062a\u0627 135 \u06cc\u0639\u0646\u06cc 45 \u0639\u0646\u0635\u0631 \u062f\u0631 \u06a9\u0644 \u062e\u0648\u0627\u0647\u062f \u0628\u0648\u062f.  \u0644\u06cc\u0633\u062a \u062f\u0648\u0645 \u0634\u0627\u0645\u0644 \u0645\u0636\u0631\u0628\u06cc \u0627\u0632 5\u060c \u0627\u0632 1 \u062a\u0627 225 \u062e\u0648\u0627\u0647\u062f \u0628\u0648\u062f. \u0644\u06cc\u0633\u062a \u062f\u0648\u0645 \u0647\u0645\u0686\u0646\u06cc\u0646 \u0634\u0627\u0645\u0644 45 \u0639\u0646\u0635\u0631 \u062f\u0631 \u06a9\u0644 \u062e\u0648\u0627\u0647\u062f \u0628\u0648\u062f.  \u0627\u0633\u06a9\u0631\u06cc\u067e\u062a \u0632\u06cc\u0631 \u0627\u06cc\u0646 \u062f\u0648 \u0644\u06cc\u0633\u062a \u0631\u0627 \u0627\u06cc\u062c\u0627\u062f \u0645\u06cc \u06a9\u0646\u062f:<\/p>\n<pre><code class=\"hljs\">X1 = np.array((x+<span class=\"hljs-number\">3<\/span> <span class=\"hljs-keyword\">for<\/span> x <span class=\"hljs-keyword\">in<\/span> <span class=\"hljs-built_in\">range<\/span>(<span class=\"hljs-number\">0<\/span>, <span class=\"hljs-number\">135<\/span>, <span class=\"hljs-number\">3<\/span>)))\n<span class=\"hljs-built_in\">print<\/span>(X1)\n\nX2 = np.array((x+<span class=\"hljs-number\">5<\/span> <span class=\"hljs-keyword\">for<\/span> x <span class=\"hljs-keyword\">in<\/span> <span class=\"hljs-built_in\">range<\/span>(<span class=\"hljs-number\">0<\/span>, <span class=\"hljs-number\">225<\/span>, <span class=\"hljs-number\">5<\/span>)))\n<span class=\"hljs-built_in\">print<\/span>(X2)\n<\/code><\/pre>\n<p>\u062f\u0631 \u062e\u0631\u0648\u062c\u06cc \u0632\u06cc\u0631 \u0645\u06cc \u062a\u0648\u0627\u0646\u06cc\u062f \u0645\u062d\u062a\u0648\u06cc\u0627\u062a \u0644\u06cc\u0633\u062a \u0631\u0627 \u0645\u0634\u0627\u0647\u062f\u0647 \u06a9\u0646\u06cc\u062f:<\/p>\n<pre><code class=\"hljs\">(  3   6   9  12  15  18  21  24  27  30  33  36  39  42  45  48  51  54\n  57  60  63  66  69  72  75  78  81  84  87  90  93  96  99 102 105 108\n 111 114 117 120 123 126 129 132 135)\n(  5  10  15  20  25  30  35  40  45  50  55  60  65  70  75  80  85  90\n  95 100 105 110 115 120 125 130 135 140 145 150 155 160 165 170 175 180\n 185 190 195 200 205 210 215 220 225)\n<\/code><\/pre>\n<p>\u0647\u0631 \u06cc\u06a9 \u0627\u0632 \u0644\u06cc\u0633\u062a \u0647\u0627\u06cc \u0628\u0627\u0644\u0627 \u0646\u0634\u0627\u0646 \u062f\u0647\u0646\u062f\u0647 \u06cc\u06a9 \u0648\u06cc\u0698\u06af\u06cc \u062f\u0631 \u0646\u0645\u0648\u0646\u0647 \u0632\u0645\u0627\u0646\u06cc \u0627\u0633\u062a.  \u0645\u062c\u0645\u0648\u0639\u0647 \u062f\u0627\u062f\u0647 \u0627\u0646\u0628\u0648\u0647 \u0631\u0627 \u0645\u06cc \u062a\u0648\u0627\u0646 \u0628\u0627 \u067e\u06cc\u0648\u0633\u062a\u0646 \u0628\u0647 \u062f\u0648 \u0644\u06cc\u0633\u062a \u0645\u0627\u0646\u0646\u062f \u0634\u06a9\u0644 \u0632\u06cc\u0631 \u0627\u06cc\u062c\u0627\u062f \u06a9\u0631\u062f:<\/p>\n<pre><code class=\"hljs\">X = np.column_stack((X1, X2))\n<span class=\"hljs-built_in\">print<\/span>(X)\n<\/code><\/pre>\n<p>\u062e\u0631\u0648\u062c\u06cc \u0645\u062c\u0645\u0648\u0639\u0647 \u062f\u0627\u062f\u0647 \u0647\u0627\u06cc \u062c\u0645\u0639 \u0622\u0648\u0631\u06cc \u0634\u062f\u0647 \u0631\u0627 \u0646\u0634\u0627\u0646 \u0645\u06cc \u062f\u0647\u062f:<\/p>\n<pre><code class=\"hljs\"> (  6  10)\n (  9  15)\n ( 12  20)\n ( 15  25)\n ( 18  30)\n ( 21  35)\n ( 24  40)\n ( 27  45)\n ( 30  50)\n ( 33  55)\n ( 36  60)\n ( 39  65)\n ( 42  70)\n ( 45  75)\n ( 48  80)\n ( 51  85)\n ( 54  90)\n ( 57  95)\n ( 60 100)\n ( 63 105)\n ( 66 110)\n ( 69 115)\n ( 72 120)\n ( 75 125)\n ( 78 130)\n ( 81 135)\n ( 84 140)\n ( 87 145)\n ( 90 150)\n ( 93 155)\n ( 96 160)\n ( 99 165)\n (102 170)\n (105 175)\n (108 180)\n (111 185)\n (114 190)\n (117 195)\n (120 200)\n (123 205)\n (126 210)\n (129 215)\n (132 220)\n (135 225))\n<\/code><\/pre>\n<p>\u0645\u0627 \u0628\u0627\u06cc\u062f \u062f\u0627\u062f\u0647 \u0647\u0627\u06cc \u062e\u0648\u062f \u0631\u0627 \u0628\u0647 \u0633\u0647 \u0628\u0639\u062f\u06cc \u062a\u063a\u06cc\u06cc\u0631 \u0634\u06a9\u0644 \u062f\u0647\u06cc\u0645 \u062a\u0627 \u0628\u062a\u0648\u0627\u0646 \u0627\u0632 \u0622\u0646\u0647\u0627 \u062a\u0648\u0633\u0637 LSTM \u0627\u0633\u062a\u0641\u0627\u062f\u0647 \u06a9\u0631\u062f.  \u0645\u0627 \u062f\u0631 \u0645\u062c\u0645\u0648\u0639 45 \u0631\u062f\u06cc\u0641 \u0648 \u062f\u0648 \u0633\u062a\u0648\u0646 \u062f\u0631 \u0645\u062c\u0645\u0648\u0639\u0647 \u062f\u0627\u062f\u0647 \u062e\u0648\u062f \u062f\u0627\u0631\u06cc\u0645.  \u0645\u0627 \u0645\u062c\u0645\u0648\u0639\u0647 \u062f\u0627\u062f\u0647 \u062e\u0648\u062f \u0631\u0627 \u0628\u0647 15 \u0646\u0645\u0648\u0646\u0647\u060c 3 \u0645\u0631\u062d\u0644\u0647 \u0632\u0645\u0627\u0646\u06cc \u0648 \u062f\u0648 \u0648\u06cc\u0698\u06af\u06cc \u062a\u063a\u06cc\u06cc\u0631 \u0634\u06a9\u0644 \u0645\u06cc \u062f\u0647\u06cc\u0645.<\/p>\n<pre><code class=\"hljs\">X = array(X).reshape(<span class=\"hljs-number\">15<\/span>, <span class=\"hljs-number\">3<\/span>, <span class=\"hljs-number\">2<\/span>)\n<span class=\"hljs-built_in\">print<\/span>(X)\n<\/code><\/pre>\n<p>\u0634\u0645\u0627 \u0645\u06cc \u062a\u0648\u0627\u0646\u06cc\u062f 15 \u0646\u0645\u0648\u0646\u0647 \u0631\u0627 \u062f\u0631 \u062e\u0631\u0648\u062c\u06cc \u0632\u06cc\u0631 \u0645\u0634\u0627\u0647\u062f\u0647 \u06a9\u0646\u06cc\u062f:<\/p>\n<pre><code class=\"hljs\">(((  3   5)\n  (  6  10)\n  (  9  15))\n\n (( 12  20)\n  ( 15  25)\n  ( 18  30))\n\n (( 21  35)\n  ( 24  40)\n  ( 27  45))\n\n (( 30  50)\n  ( 33  55)\n  ( 36  60))\n\n (( 39  65)\n  ( 42  70)\n  ( 45  75))\n\n (( 48  80)\n  ( 51  85)\n  ( 54  90))\n\n (( 57  95)\n  ( 60 100)\n  ( 63 105))\n\n (( 66 110)\n  ( 69 115)\n  ( 72 120))\n\n (( 75 125)\n  ( 78 130)\n  ( 81 135))\n\n (( 84 140)\n  ( 87 145)\n  ( 90 150))\n\n (( 93 155)\n  ( 96 160)\n  ( 99 165))\n\n ((102 170)\n  (105 175)\n  (108 180))\n\n ((111 185)\n  (114 190)\n  (117 195))\n\n ((120 200)\n  (123 205)\n  (126 210))\n\n ((129 215)\n  (132 220)\n  (135 225)))\n<\/code><\/pre>\n<p>\u062e\u0631\u0648\u062c\u06cc \u0646\u06cc\u0632 \u062f\u0627\u0631\u0627\u06cc 15 \u0645\u0642\u062f\u0627\u0631 \u0645\u0631\u0628\u0648\u0637 \u0628\u0647 15 \u0646\u0645\u0648\u0646\u0647 \u0648\u0631\u0648\u062f\u06cc \u062e\u0648\u0627\u0647\u062f \u0628\u0648\u062f.  \u0647\u0631 \u0645\u0642\u062f\u0627\u0631 \u062f\u0631 \u062e\u0631\u0648\u062c\u06cc\u060c \u0645\u062c\u0645\u0648\u0639 \u062f\u0648 \u0645\u0642\u062f\u0627\u0631 \u0645\u0634\u062e\u0635\u0647 \u062f\u0631 \u0633\u0648\u0645\u06cc\u0646 \u0645\u0631\u062d\u0644\u0647 \u0632\u0645\u0627\u0646\u06cc \u0647\u0631 \u0646\u0645\u0648\u0646\u0647 \u0648\u0631\u0648\u062f\u06cc \u062e\u0648\u0627\u0647\u062f \u0628\u0648\u062f.  \u0628\u0647 \u0639\u0646\u0648\u0627\u0646 \u0645\u062b\u0627\u0644\u060c \u0633\u0648\u0645\u06cc\u0646 \u0645\u0631\u062d\u0644\u0647 \u0632\u0645\u0627\u0646\u06cc \u0646\u0645\u0648\u0646\u0647 \u0627\u0648\u0644 \u062f\u0627\u0631\u0627\u06cc \u0648\u06cc\u0698\u06af\u06cc \u0647\u0627\u06cc 9 \u0648 15 \u0627\u0633\u062a\u060c \u0628\u0646\u0627\u0628\u0631\u0627\u06cc\u0646 \u062e\u0631\u0648\u062c\u06cc 24 \u062e\u0648\u0627\u0647\u062f \u0628\u0648\u062f. \u0628\u0647 \u0637\u0648\u0631 \u0645\u0634\u0627\u0628\u0647\u060c \u062f\u0648 \u0645\u0642\u062f\u0627\u0631 \u0648\u06cc\u0698\u06af\u06cc \u062f\u0631 \u0645\u0631\u062d\u0644\u0647 \u0632\u0645\u0627\u0646\u06cc \u0633\u0648\u0645 \u0646\u0645\u0648\u0646\u0647 \u062f\u0648\u0645 18 \u0648 30 \u0647\u0633\u062a\u0646\u062f.  \u062e\u0631\u0648\u062c\u06cc \u0645\u0631\u0628\u0648\u0637\u0647 48 \u062e\u0648\u0627\u0647\u062f \u0628\u0648\u062f \u0648 \u0628\u0647 \u0647\u0645\u06cc\u0646 \u062a\u0631\u062a\u06cc\u0628 \u0631\u0648\u06cc.<\/p>\n<p>\u0627\u0633\u06a9\u0631\u06cc\u067e\u062a \u0632\u06cc\u0631 \u0628\u0631\u062f\u0627\u0631 \u062e\u0631\u0648\u062c\u06cc \u0631\u0627 \u0627\u06cc\u062c\u0627\u062f \u0648 \u0646\u0645\u0627\u06cc\u0634 \u0645\u06cc \u062f\u0647\u062f:<\/p>\n<pre><code class=\"hljs\">( 24  48  72  96 120 144 168 192 216 240 264 288 312 336 360)\n<\/code><\/pre>\n<p>\u0628\u06cc\u0627\u06cc\u06cc\u062f \u0627\u06a9\u0646\u0648\u0646 \u0627\u06cc\u0646 \u0645\u0634\u06a9\u0644 \u062a\u0648\u0627\u0644\u06cc \u0686\u0646\u062f \u0628\u0647 \u06cc\u06a9 \u0631\u0627 \u0627\u0632 \u0637\u0631\u06cc\u0642 LSTM\u0647\u0627\u06cc \u0633\u0627\u062f\u0647\u060c \u0627\u0646\u0628\u0627\u0634\u062a\u0647 \u0648 \u062f\u0648 \u0637\u0631\u0641\u0647 \u062d\u0644 \u06a9\u0646\u06cc\u0645.<\/p>\n<h4 id=\"solutionviasimplelstm\">\u0631\u0627\u0647 \u062d\u0644 \u0627\u0632 \u0637\u0631\u06cc\u0642 LSTM \u0633\u0627\u062f\u0647<\/h4>\n<pre><code class=\"hljs\">model = Sequential()\nmodel.add(LSTM(<span class=\"hljs-number\">50<\/span>, activation=<span class=\"hljs-string\">'relu'<\/span>, input_shape=(<span class=\"hljs-number\">3<\/span>, <span class=\"hljs-number\">2<\/span>)))\nmodel.add(Dense(<span class=\"hljs-number\">1<\/span>))\nmodel.<span class=\"hljs-built_in\">compile<\/span>(optimizer=<span class=\"hljs-string\">'adam'<\/span>, loss=<span class=\"hljs-string\">'mse'<\/span>)\nhistory = model.fit(X, Y, epochs=<span class=\"hljs-number\">1000<\/span>, validation_split=<span class=\"hljs-number\">0.2<\/span>, verbose=<span class=\"hljs-number\">1<\/span>)\n<\/code><\/pre>\n<p>\u0645\u062f\u0644 \u0622\u0645\u0648\u0632\u0634 \u062f\u06cc\u062f\u0647 \u0627\u0633\u062a.  \u0645\u0627 \u06cc\u06a9 \u0646\u0642\u0637\u0647 \u062f\u0627\u062f\u0647 \u0622\u0632\u0645\u0627\u06cc\u0634\u06cc \u0627\u06cc\u062c\u0627\u062f \u0645\u06cc \u06a9\u0646\u06cc\u0645 \u0648 \u0633\u067e\u0633 \u0627\u0632 \u0645\u062f\u0644 \u062e\u0648\u062f \u0628\u0631\u0627\u06cc \u067e\u06cc\u0634 \u0628\u06cc\u0646\u06cc \u0627\u0633\u062a\u0641\u0627\u062f\u0647 \u0645\u06cc \u06a9\u0646\u06cc\u0645 \u0631\u0648\u06cc \u0646\u0642\u0637\u0647 \u0622\u0632\u0645\u0648\u0646<\/p>\n<pre><code class=\"hljs\">test_input = array(((<span class=\"hljs-number\">8<\/span>, <span class=\"hljs-number\">51<\/span>),\n                    (<span class=\"hljs-number\">11<\/span>,<span class=\"hljs-number\">56<\/span>),\n                    (<span class=\"hljs-number\">14<\/span>,<span class=\"hljs-number\">61<\/span>)))\n\ntest_input = test_input.reshape((<span class=\"hljs-number\">1<\/span>, <span class=\"hljs-number\">3<\/span>, <span class=\"hljs-number\">2<\/span>))\ntest_output = model.predict(test_input, verbose=<span class=\"hljs-number\">0<\/span>)\n<span class=\"hljs-built_in\">print<\/span>(test_output)\n<\/code><\/pre>\n<p>\u0645\u062c\u0645\u0648\u0639 \u062f\u0648 \u0648\u06cc\u0698\u06af\u06cc \u0633\u0648\u0645\u06cc\u0646 \u0645\u0631\u062d\u0644\u0647 \u0632\u0645\u0627\u0646\u06cc \u0648\u0631\u0648\u062f\u06cc 14 + 61 = 75 \u0627\u0633\u062a. \u0645\u062f\u0644 \u0645\u0627 \u0628\u0627 \u06cc\u06a9 \u0644\u0627\u06cc\u0647 LSTM 73.41 \u0631\u0627 \u067e\u06cc\u0634 \u0628\u06cc\u0646\u06cc \u06a9\u0631\u062f \u06a9\u0647 \u0628\u0633\u06cc\u0627\u0631 \u0646\u0632\u062f\u06cc\u06a9 \u0627\u0633\u062a.<\/p>\n<h4 id=\"solutionviastackedlstm\">\u0631\u0627\u0647 \u062d\u0644 \u0627\u0632 \u0637\u0631\u06cc\u0642 Stacked LSTM<\/h4>\n<p>\u0627\u0633\u06a9\u0631\u06cc\u067e\u062a \u0632\u06cc\u0631 \u06cc\u06a9 LSTM \u0627\u0646\u0628\u0627\u0634\u062a\u0647 \u0631\u0627 \u0622\u0645\u0648\u0632\u0634 \u0645\u06cc \u062f\u0647\u062f \u0648 \u067e\u06cc\u0634 \u0628\u06cc\u0646\u06cc \u0645\u06cc \u06a9\u0646\u062f \u0631\u0648\u06cc \u0646\u0642\u0637\u0647 \u06cc \u0627\u0645\u062a\u062d\u0627\u0646:<\/p>\n<pre><code class=\"hljs\">model = Sequential()\nmodel.add(LSTM(<span class=\"hljs-number\">200<\/span>, activation=<span class=\"hljs-string\">'relu'<\/span>, return_sequences=<span class=\"hljs-literal\">True<\/span>, input_shape=(<span class=\"hljs-number\">3<\/span>, <span class=\"hljs-number\">2<\/span>)))\nmodel.add(LSTM(<span class=\"hljs-number\">100<\/span>, activation=<span class=\"hljs-string\">'relu'<\/span>, return_sequences=<span class=\"hljs-literal\">True<\/span>))\nmodel.add(LSTM(<span class=\"hljs-number\">50<\/span>, activation=<span class=\"hljs-string\">'relu'<\/span>, return_sequences=<span class=\"hljs-literal\">True<\/span>))\nmodel.add(LSTM(<span class=\"hljs-number\">25<\/span>, activation=<span class=\"hljs-string\">'relu'<\/span>))\nmodel.add(Dense(<span class=\"hljs-number\">20<\/span>, activation=<span class=\"hljs-string\">'relu'<\/span>))\nmodel.add(Dense(<span class=\"hljs-number\">10<\/span>, activation=<span class=\"hljs-string\">'relu'<\/span>))\nmodel.add(Dense(<span class=\"hljs-number\">1<\/span>))\nmodel.<span class=\"hljs-built_in\">compile<\/span>(optimizer=<span class=\"hljs-string\">'adam'<\/span>, loss=<span class=\"hljs-string\">'mse'<\/span>)\n\nhistory = model.fit(X, Y, epochs=<span class=\"hljs-number\">1000<\/span>, validation_split=<span class=\"hljs-number\">0.2<\/span>, verbose=<span class=\"hljs-number\">1<\/span>)\n\ntest_output = model.predict(test_input, verbose=<span class=\"hljs-number\">0<\/span>)\n<span class=\"hljs-built_in\">print<\/span>(test_output)\n<\/code><\/pre>\n<p>\u062e\u0631\u0648\u062c\u06cc \u0645\u0646 71.56 \u0627\u0633\u062a \u06a9\u0647 \u0627\u0632 LSTM \u0633\u0627\u062f\u0647 \u0628\u062f\u062a\u0631 \u0627\u0633\u062a.  \u0628\u0647 \u0646\u0638\u0631 \u0645\u06cc \u0631\u0633\u062f LSTM \u0627\u0646\u0628\u0627\u0634\u062a\u0647 \u0645\u0627 \u0628\u06cc\u0634 \u0627\u0632 \u062d\u062f \u0645\u0646\u0627\u0633\u0628 \u0627\u0633\u062a.<\/p>\n<h4 id=\"solutionviabidirectionallstm\">\u0631\u0627\u0647 \u062d\u0644 \u0627\u0632 \u0637\u0631\u06cc\u0642 \u062f\u0648 \u062c\u0647\u062a\u0647 LSTM<\/h4>\n<p>\u062f\u0631 \u0627\u06cc\u0646\u062c\u0627 \u0627\u0633\u06a9\u0631\u06cc\u067e\u062a \u0622\u0645\u0648\u0632\u0634\u06cc \u0628\u0631\u0627\u06cc LSTM \u062f\u0648 \u062c\u0647\u062a\u0647 \u0633\u0627\u062f\u0647 \u0628\u0647 \u0647\u0645\u0631\u0627\u0647 \u06a9\u062f\u06cc \u06a9\u0647 \u0628\u0631\u0627\u06cc \u067e\u06cc\u0634 \u0628\u06cc\u0646\u06cc \u0627\u0633\u062a\u0641\u0627\u062f\u0647 \u0645\u06cc \u0634\u0648\u062f\u060c \u0622\u0645\u062f\u0647 \u0627\u0633\u062a \u0631\u0648\u06cc \u0646\u0642\u0637\u0647 \u062f\u0627\u062f\u0647 \u062a\u0633\u062a:<\/p>\n<pre><code class=\"hljs\"><span class=\"hljs-keyword\">from<\/span> keras.layers <span class=\"hljs-keyword\">import<\/span> Bidirectional\n\nmodel = Sequential()\nmodel.add(Bidirectional(LSTM(<span class=\"hljs-number\">50<\/span>, activation=<span class=\"hljs-string\">'relu'<\/span>), input_shape=(<span class=\"hljs-number\">3<\/span>, <span class=\"hljs-number\">2<\/span>)))\nmodel.add(Dense(<span class=\"hljs-number\">1<\/span>))\nmodel.<span class=\"hljs-built_in\">compile<\/span>(optimizer=<span class=\"hljs-string\">'adam'<\/span>, loss=<span class=\"hljs-string\">'mse'<\/span>)\n\nhistory = model.fit(X, Y, epochs=<span class=\"hljs-number\">1000<\/span>, validation_split=<span class=\"hljs-number\">0.2<\/span>, verbose=<span class=\"hljs-number\">1<\/span>)\ntest_output = model.predict(test_input, verbose=<span class=\"hljs-number\">0<\/span>)\n<span class=\"hljs-built_in\">print<\/span>(test_output)\n<\/code><\/pre>\n<p>\u062e\u0631\u0648\u062c\u06cc 76.82 \u0627\u0633\u062a \u06a9\u0647 \u062a\u0642\u0631\u06cc\u0628\u0627\u064b \u0646\u0632\u062f\u06cc\u06a9 \u0628\u0647 75 \u0627\u0633\u062a. \u0628\u0627\u0632 \u0647\u0645 \u0628\u0647 \u0646\u0638\u0631 \u0645\u06cc \u0631\u0633\u062f LSTM \u062f\u0648 \u0637\u0631\u0641\u0647 \u0627\u0632 \u0628\u0642\u06cc\u0647 \u0627\u0644\u06af\u0648\u0631\u06cc\u062a\u0645 \u0647\u0627 \u0628\u0647\u062a\u0631 \u0639\u0645\u0644 \u0645\u06cc \u06a9\u0646\u062f.<\/p>\n<p>\u062a\u0627 \u06a9\u0646\u0648\u0646 \u0645\u0627 \u0628\u0631 \u0627\u0633\u0627\u0633 \u0645\u0642\u0627\u062f\u06cc\u0631 \u062a\u06a9 \u067e\u06cc\u0634\u200c\u0628\u06cc\u0646\u06cc \u06a9\u0631\u062f\u0647\u200c\u0627\u06cc\u0645 \u0631\u0648\u06cc \u0645\u0642\u0627\u062f\u06cc\u0631 \u0686\u0646\u062f \u0648\u06cc\u0698\u06af\u06cc \u0627\u0632 \u0645\u0631\u0627\u062d\u0644 \u0632\u0645\u0627\u0646\u06cc \u0645\u062e\u062a\u0644\u0641.  \u0645\u0648\u0631\u062f \u062f\u06cc\u06af\u0631\u06cc \u0627\u0632 \u062a\u0648\u0627\u0644\u06cc \u0647\u0627\u06cc \u0686\u0646\u062f \u0628\u0647 \u06cc\u06a9 \u0648\u062c\u0648\u062f \u062f\u0627\u0631\u062f \u06a9\u0647 \u062f\u0631 \u0622\u0646 \u0645\u06cc \u062e\u0648\u0627\u0647\u06cc\u062f \u06cc\u06a9 \u0645\u0642\u062f\u0627\u0631 \u0628\u0631\u0627\u06cc \u0647\u0631 \u0648\u06cc\u0698\u06af\u06cc \u062f\u0631 \u0645\u0631\u062d\u0644\u0647 \u0632\u0645\u0627\u0646\u06cc \u067e\u06cc\u0634 \u0628\u06cc\u0646\u06cc \u06a9\u0646\u06cc\u062f.  \u0628\u0647 \u0639\u0646\u0648\u0627\u0646 \u0645\u062b\u0627\u0644\u060c \u0645\u062c\u0645\u0648\u0639\u0647 \u062f\u0627\u062f\u0647 \u0627\u06cc \u06a9\u0647 \u0645\u0627 \u062f\u0631 \u0627\u06cc\u0646 \u0628\u062e\u0634 \u0627\u0633\u062a\u0641\u0627\u062f\u0647 \u06a9\u0631\u062f\u06cc\u0645 \u062f\u0627\u0631\u0627\u06cc \u0633\u0647 \u0645\u0631\u062d\u0644\u0647 \u0632\u0645\u0627\u0646\u06cc \u0627\u0633\u062a \u0648 \u0647\u0631 \u0645\u0631\u062d\u0644\u0647 \u0632\u0645\u0627\u0646\u06cc \u062f\u0627\u0631\u0627\u06cc \u062f\u0648 \u0648\u06cc\u0698\u06af\u06cc \u0627\u0633\u062a.  \u0645\u0645\u06a9\u0646 \u0627\u0633\u062a \u0628\u062e\u0648\u0627\u0647\u06cc\u0645 \u0645\u0642\u0627\u062f\u06cc\u0631 \u0641\u0631\u062f\u06cc \u0631\u0627 \u0628\u0631\u0627\u06cc \u0647\u0631 \u0633\u0631\u06cc \u0648\u06cc\u0698\u06af\u06cc \u067e\u06cc\u0634 \u0628\u06cc\u0646\u06cc \u06a9\u0646\u06cc\u0645.  \u0645\u062b\u0627\u0644 \u0632\u06cc\u0631 \u0631\u0648\u0634\u0646 \u0645\u06cc \u06a9\u0646\u062f\u060c \u0641\u0631\u0636 \u06a9\u0646\u06cc\u062f \u0648\u0631\u0648\u062f\u06cc \u0632\u06cc\u0631 \u0631\u0627 \u062f\u0627\u0631\u06cc\u0645:<\/p>\n<pre><code class=\"hljs\">(((  3   5)\n  (  6  10)\n  (  9  15))\n<\/code><\/pre>\n<p>\u062f\u0631 \u062e\u0631\u0648\u062c\u06cc\u060c \u0645\u0627 \u06cc\u06a9 \u0645\u0631\u062d\u0644\u0647 \u0632\u0645\u0627\u0646\u06cc \u0628\u0627 \u062f\u0648 \u0648\u06cc\u0698\u06af\u06cc \u0645\u0627\u0646\u0646\u062f \u0634\u06a9\u0644 \u0632\u06cc\u0631 \u0645\u06cc \u062e\u0648\u0627\u0647\u06cc\u0645:<\/p>\n<pre><code class=\"hljs\">(12, 20)\n<\/code><\/pre>\n<p>\u0645\u06cc \u0628\u06cc\u0646\u06cc\u062f \u06a9\u0647 \u0645\u0642\u062f\u0627\u0631 \u0627\u0648\u0644 \u062f\u0631 \u062e\u0631\u0648\u062c\u06cc \u0627\u062f\u0627\u0645\u0647 \u0633\u0631\u06cc \u0627\u0648\u0644 \u0648 \u0645\u0642\u062f\u0627\u0631 \u062f\u0648\u0645 \u0627\u062f\u0627\u0645\u0647 \u0633\u0631\u06cc \u062f\u0648\u0645 \u0627\u0633\u062a.  \u0645\u0627 \u0645\u06cc\u200c\u062a\u0648\u0627\u0646\u06cc\u0645 \u0686\u0646\u06cc\u0646 \u0645\u0634\u06a9\u0644\u0627\u062a\u06cc \u0631\u0627 \u0628\u0627 \u062a\u063a\u06cc\u06cc\u0631 \u062a\u0639\u062f\u0627\u062f \u0646\u0648\u0631\u0648\u0646\u200c\u0647\u0627 \u062f\u0631 \u0644\u0627\u06cc\u0647 \u0645\u062a\u0631\u0627\u06a9\u0645 \u062e\u0631\u0648\u062c\u06cc \u0628\u0647 \u062a\u0639\u062f\u0627\u062f \u0645\u0642\u0627\u062f\u06cc\u0631 \u0648\u06cc\u0698\u06af\u06cc\u200c\u0647\u0627\u06cc\u06cc \u06a9\u0647 \u062f\u0631 \u062e\u0631\u0648\u062c\u06cc \u0645\u06cc\u200c\u062e\u0648\u0627\u0647\u06cc\u0645 \u062d\u0644 \u06a9\u0646\u06cc\u0645.  \u0628\u0627 \u0627\u06cc\u0646 \u062d\u0627\u0644\u060c \u0627\u0628\u062a\u062f\u0627 \u0628\u0627\u06cc\u062f \u0628\u0631\u062f\u0627\u0631 \u062e\u0631\u0648\u062c\u06cc \u062e\u0648\u062f \u0631\u0627 \u0628\u0647 \u0631\u0648\u0632 \u06a9\u0646\u06cc\u0645 <code>Y<\/code>.  \u0628\u0631\u062f\u0627\u0631 \u0648\u0631\u0648\u062f\u06cc \u062b\u0627\u0628\u062a \u0645\u06cc \u0645\u0627\u0646\u062f:<\/p>\n<pre><code class=\"hljs\">Y = <span class=\"hljs-built_in\">list<\/span>()\n<span class=\"hljs-keyword\">for<\/span> x <span class=\"hljs-keyword\">in<\/span> X:\n    new_item = <span class=\"hljs-built_in\">list<\/span>()\n    new_item.append(x(<span class=\"hljs-number\">2<\/span>)(<span class=\"hljs-number\">0<\/span>)+<span class=\"hljs-number\">3<\/span>)\n    new_item.append(x(<span class=\"hljs-number\">2<\/span>)(<span class=\"hljs-number\">1<\/span>)+<span class=\"hljs-number\">5<\/span>)\n    Y.append(new_item)\n\nY = np.array(Y)\n<span class=\"hljs-built_in\">print<\/span>(Y)\n<\/code><\/pre>\n<p>\u0627\u0633\u06a9\u0631\u06cc\u067e\u062a \u0628\u0627\u0644\u0627 \u06cc\u06a9 \u0628\u0631\u062f\u0627\u0631 \u062e\u0631\u0648\u062c\u06cc \u0628\u0647 \u0631\u0648\u0632 \u0634\u062f\u0647 \u0627\u06cc\u062c\u0627\u062f \u0645\u06cc \u06a9\u0646\u062f \u0648 \u0622\u0646 \u0631\u0627 \u0686\u0627\u067e \u0645\u06cc \u06a9\u0646\u062f \u0631\u0648\u06cc \u0631\u0627 console\u060c \u062e\u0631\u0648\u062c\u06cc \u0628\u0647 \u0635\u0648\u0631\u062a \u0632\u06cc\u0631 \u0627\u0633\u062a:<\/p>\n<pre><code class=\"hljs\">(( 12  20)\n ( 21  35)\n ( 30  50)\n ( 39  65)\n ( 48  80)\n ( 57  95)\n ( 66 110)\n ( 75 125)\n ( 84 140)\n ( 93 155)\n (102 170)\n (111 185)\n (120 200)\n (129 215)\n (138 230))\n<\/code><\/pre>\n<p>\u0628\u06cc\u0627\u06cc\u06cc\u062f \u0627\u06a9\u0646\u0648\u0646 \u0634\u0628\u06a9\u0647 \u0647\u0627\u06cc LSTM \u0633\u0627\u062f\u0647\u060c \u067e\u0634\u062a\u0647 \u0627\u06cc \u0648 \u062f\u0648 \u0637\u0631\u0641\u0647 \u062e\u0648\u062f \u0631\u0627 \u0622\u0645\u0648\u0632\u0634 \u062f\u0647\u06cc\u0645 \u0631\u0648\u06cc \u0645\u062c\u0645\u0648\u0639\u0647 \u062f\u0627\u062f\u0647 \u0645\u0627  \u0627\u0633\u06a9\u0631\u06cc\u067e\u062a \u0632\u06cc\u0631 \u06cc\u06a9 LSTM \u0633\u0627\u062f\u0647 \u0631\u0627 \u0622\u0645\u0648\u0632\u0634 \u0645\u06cc \u062f\u0647\u062f:<\/p>\n<pre><code class=\"hljs\">model = Sequential()\nmodel.add(LSTM(<span class=\"hljs-number\">50<\/span>, activation=<span class=\"hljs-string\">'relu'<\/span>, input_shape=(<span class=\"hljs-number\">3<\/span>, <span class=\"hljs-number\">2<\/span>)))\nmodel.add(Dense(<span class=\"hljs-number\">2<\/span>))\nmodel.<span class=\"hljs-built_in\">compile<\/span>(optimizer=<span class=\"hljs-string\">'adam'<\/span>, loss=<span class=\"hljs-string\">'mse'<\/span>)\n\nhistory = model.fit(X, Y, epochs=<span class=\"hljs-number\">1000<\/span>, validation_split=<span class=\"hljs-number\">0.2<\/span>, verbose=<span class=\"hljs-number\">1<\/span>)\n<\/code><\/pre>\n<p>\u06af\u0627\u0645 \u0628\u0639\u062f\u06cc \u0627\u06cc\u0646 \u0627\u0633\u062a \u06a9\u0647 \u0645\u062f\u0644 \u062e\u0648\u062f \u0631\u0627 \u0622\u0632\u0645\u0627\u06cc\u0634 \u06a9\u0646\u06cc\u0645 \u0631\u0648\u06cc \u0646\u0642\u0637\u0647 \u062f\u0627\u062f\u0647 \u062a\u0633\u062a  \u0627\u0633\u06a9\u0631\u06cc\u067e\u062a \u0632\u06cc\u0631 \u06cc\u06a9 \u0646\u0642\u0637\u0647 \u062f\u0627\u062f\u0647 \u0622\u0632\u0645\u0627\u06cc\u0634\u06cc \u0627\u06cc\u062c\u0627\u062f \u0645\u06cc \u06a9\u0646\u062f:<\/p>\n<pre><code class=\"hljs\">test_input = array(((<span class=\"hljs-number\">20<\/span>,<span class=\"hljs-number\">34<\/span>),\n                    (<span class=\"hljs-number\">23<\/span>,<span class=\"hljs-number\">39<\/span>),\n                    (<span class=\"hljs-number\">26<\/span>,<span class=\"hljs-number\">44<\/span>)))\n\ntest_input = test_input.reshape((<span class=\"hljs-number\">1<\/span>, <span class=\"hljs-number\">3<\/span>, <span class=\"hljs-number\">2<\/span>))\ntest_output = model.predict(test_input, verbose=<span class=\"hljs-number\">0<\/span>)\n<span class=\"hljs-built_in\">print<\/span>(test_output)\n<\/code><\/pre>\n<p>\u062e\u0631\u0648\u062c\u06cc \u0648\u0627\u0642\u0639\u06cc (29\u060c 45) \u0627\u0633\u062a.  \u0645\u062f\u0644 \u0645\u0627 (29.089157\u060c 48.469097) \u0631\u0627 \u067e\u06cc\u0634 \u0628\u06cc\u0646\u06cc \u0645\u06cc \u06a9\u0646\u062f \u06a9\u0647 \u0628\u0633\u06cc\u0627\u0631 \u0646\u0632\u062f\u06cc\u06a9 \u0627\u0633\u062a.<\/p>\n<p>\u0628\u06cc\u0627\u06cc\u06cc\u062f \u0627\u06a9\u0646\u0648\u0646 \u06cc\u06a9 LSTM \u0627\u0646\u0628\u0627\u0634\u062a\u0647 \u0631\u0627 \u0622\u0645\u0648\u0632\u0634 \u062f\u0647\u06cc\u0645 \u0648 \u062e\u0631\u0648\u062c\u06cc \u0631\u0627 \u0628\u0631\u0627\u06cc \u0646\u0642\u0637\u0647 \u062f\u0627\u062f\u0647 \u0622\u0632\u0645\u0627\u06cc\u0634\u06cc \u067e\u06cc\u0634 \u0628\u06cc\u0646\u06cc \u06a9\u0646\u06cc\u0645:<\/p>\n<pre><code class=\"hljs\">model = Sequential()\nmodel.add(LSTM(<span class=\"hljs-number\">100<\/span>, activation=<span class=\"hljs-string\">'relu'<\/span>, return_sequences=<span class=\"hljs-literal\">True<\/span>, input_shape=(<span class=\"hljs-number\">3<\/span>, <span class=\"hljs-number\">2<\/span>)))\nmodel.add(LSTM(<span class=\"hljs-number\">50<\/span>, activation=<span class=\"hljs-string\">'relu'<\/span>, return_sequences=<span class=\"hljs-literal\">True<\/span>))\nmodel.add(LSTM(<span class=\"hljs-number\">25<\/span>, activation=<span class=\"hljs-string\">'relu'<\/span>))\nmodel.add(Dense(<span class=\"hljs-number\">10<\/span>, activation=<span class=\"hljs-string\">'relu'<\/span>))\nmodel.add(Dense(<span class=\"hljs-number\">2<\/span>))\nmodel.<span class=\"hljs-built_in\">compile<\/span>(optimizer=<span class=\"hljs-string\">'adam'<\/span>, loss=<span class=\"hljs-string\">'mse'<\/span>)\n\nhistory = model.fit(X, Y, epochs=<span class=\"hljs-number\">500<\/span>, validation_split=<span class=\"hljs-number\">0.2<\/span>, verbose=<span class=\"hljs-number\">1<\/span>)\n\ntest_output = model.predict(test_input, verbose=<span class=\"hljs-number\">0<\/span>)\n<span class=\"hljs-built_in\">print<\/span>(test_output)\n<\/code><\/pre>\n<p>\u062e\u0631\u0648\u062c\u06cc (29.170143\u060c 48.688267) \u0627\u0633\u062a \u06a9\u0647 \u0628\u0627\u0632 \u0647\u0645 \u0628\u0633\u06cc\u0627\u0631 \u0646\u0632\u062f\u06cc\u06a9 \u0628\u0647 \u062e\u0631\u0648\u062c\u06cc \u0648\u0627\u0642\u0639\u06cc \u0627\u0633\u062a.<\/p>\n<p>\u062f\u0631 \u0646\u0647\u0627\u06cc\u062a\u060c \u0645\u0627 \u0645\u06cc \u062a\u0648\u0627\u0646\u06cc\u0645 LSTM \u062f\u0648 \u0637\u0631\u0641\u0647 \u062e\u0648\u062f \u0631\u0627 \u0622\u0645\u0648\u0632\u0634 \u062f\u0647\u06cc\u0645 \u0648 \u067e\u06cc\u0634 \u0628\u06cc\u0646\u06cc \u06a9\u0646\u06cc\u0645 \u0631\u0648\u06cc \u0646\u0642\u0637\u0647 \u0622\u0632\u0645\u0648\u0646:<\/p>\n<pre><code class=\"hljs\"><span class=\"hljs-keyword\">from<\/span> keras.layers <span class=\"hljs-keyword\">import<\/span> Bidirectional\n\nmodel = Sequential()\nmodel.add(Bidirectional(LSTM(<span class=\"hljs-number\">50<\/span>, activation=<span class=\"hljs-string\">'relu'<\/span>), input_shape=(<span class=\"hljs-number\">3<\/span>, <span class=\"hljs-number\">2<\/span>)))\nmodel.add(Dense(<span class=\"hljs-number\">2<\/span>))\nmodel.<span class=\"hljs-built_in\">compile<\/span>(optimizer=<span class=\"hljs-string\">'adam'<\/span>, loss=<span class=\"hljs-string\">'mse'<\/span>)\n\nhistory = model.fit(X, Y, epochs=<span class=\"hljs-number\">1000<\/span>, validation_split=<span class=\"hljs-number\">0.2<\/span>, verbose=<span class=\"hljs-number\">1<\/span>)\ntest_output = model.predict(test_input, verbose=<span class=\"hljs-number\">0<\/span>)\n<span class=\"hljs-built_in\">print<\/span>(test_output)\n<\/code><\/pre>\n<p>\u062e\u0631\u0648\u062c\u06cc (29.2071\u060c 48.737988) \u0627\u0633\u062a.<\/p>\n<p>\u0645\u06cc \u062a\u0648\u0627\u0646\u06cc\u062f \u06cc\u06a9 \u0628\u0627\u0631 \u062f\u06cc\u06af\u0631 \u0628\u0628\u06cc\u0646\u06cc\u062f \u06a9\u0647 LSTM \u062f\u0648 \u0637\u0631\u0641\u0647 \u062f\u0642\u06cc\u0642 \u062a\u0631\u06cc\u0646 \u067e\u06cc\u0634 \u0628\u06cc\u0646\u06cc \u0631\u0627 \u0627\u0646\u062c\u0627\u0645 \u0645\u06cc \u062f\u0647\u062f.<\/p>\n<h2 id=\"conclusion\"><span class=\"ez-toc-section\" id=\"%d9%86%d8%aa%db%8c%d8%ac%d9%87\"><\/span>\u0646\u062a\u06cc\u062c\u0647<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<p>\u0634\u0628\u06a9\u0647 \u0647\u0627\u06cc \u0639\u0635\u0628\u06cc \u0633\u0627\u062f\u0647 \u0628\u0631\u0627\u06cc \u062d\u0644 \u0645\u0633\u0627\u0626\u0644 \u062a\u0648\u0627\u0644\u06cc \u0645\u0646\u0627\u0633\u0628 \u0646\u06cc\u0633\u062a\u0646\u062f \u0632\u06cc\u0631\u0627 \u062f\u0631 \u0645\u0633\u0627\u0626\u0644 \u062a\u0648\u0627\u0644\u06cc \u0639\u0644\u0627\u0648\u0647 \u0628\u0631 \u0648\u0631\u0648\u062f\u06cc \u0641\u0639\u0644\u06cc\u060c \u0628\u0627\u06cc\u062f \u0648\u0631\u0648\u062f\u06cc \u0647\u0627\u06cc \u0642\u0628\u0644\u06cc \u0631\u0627 \u0646\u06cc\u0632 \u067e\u06cc\u06af\u06cc\u0631\u06cc \u06a9\u0646\u06cc\u0645.  \u0634\u0628\u06a9\u0647 \u0647\u0627\u06cc \u0639\u0635\u0628\u06cc \u0628\u0627 \u0646\u0648\u0639\u06cc \u062d\u0627\u0641\u0638\u0647 \u0628\u0631\u0627\u06cc \u062d\u0644 \u0645\u0633\u0627\u0626\u0644 \u062a\u0648\u0627\u0644\u06cc \u0645\u0646\u0627\u0633\u0628 \u062a\u0631 \u0647\u0633\u062a\u0646\u062f.  LSTM \u06cc\u06a9\u06cc \u0627\u0632 \u0627\u06cc\u0646 \u0634\u0628\u06a9\u0647 \u0647\u0627\u0633\u062a.<\/p>\n<p>\u062f\u0631 \u0627\u06cc\u0646 \u0645\u0642\u0627\u0644\u0647 \u062f\u06cc\u062f\u06cc\u0645 \u06a9\u0647 \u0686\u06af\u0648\u0646\u0647 \u0645\u06cc \u062a\u0648\u0627\u0646 \u0627\u0632 \u0627\u0646\u0648\u0627\u0639 \u0645\u062e\u062a\u0644\u0641 \u0627\u0644\u06af\u0648\u0631\u06cc\u062a\u0645 LSTM \u0628\u0631\u0627\u06cc \u062d\u0644 \u0645\u0633\u0627\u0626\u0644 \u062a\u0648\u0627\u0644\u06cc \u06cc\u06a9 \u0628\u0647 \u06cc\u06a9 \u0648 \u0686\u0646\u062f \u0628\u0647 \u06cc\u06a9 \u0627\u0633\u062a\u0641\u0627\u062f\u0647 \u06a9\u0631\u062f.  \u0627\u06cc\u0646 \u0642\u0633\u0645\u062a \u0627\u0648\u0644 \u0645\u0642\u0627\u0644\u0647 \u0627\u0633\u062a.  \u062f\u0631 \u0642\u0633\u0645\u062a \u062f\u0648\u0645 \u0631\u0648\u0634 \u062d\u0644 \u0645\u0633\u0627\u0626\u0644 \u062a\u0648\u0627\u0644\u06cc \u06cc\u06a9 \u0628\u0647 \u0686\u0646\u062f \u0648 \u0686\u0646\u062f \u0628\u0647 \u0686\u0646\u062f \u0631\u0627 \u062e\u0648\u0627\u0647\u06cc\u0645 \u062f\u06cc\u062f.  \u0647\u0645\u0686\u0646\u06cc\u0646 \u0645\u06a9\u0627\u0646\u06cc\u0633\u0645 \u0631\u0645\u0632\u06af\u0630\u0627\u0631-\u0631\u0645\u0632\u06af\u0634\u0627\u06cc\u06cc \u0631\u0627 \u06a9\u0647 \u0628\u06cc\u0634\u062a\u0631 \u0628\u0631\u0627\u06cc \u0627\u06cc\u062c\u0627\u062f \u0631\u0628\u0627\u062a\u200c\u0647\u0627\u06cc \u06af\u0641\u062a\u06af\u0648 \u0627\u0633\u062a\u0641\u0627\u062f\u0647 \u0645\u06cc\u200c\u0634\u0648\u062f\u060c \u0645\u0637\u0627\u0644\u0639\u0647 \u062e\u0648\u0627\u0647\u06cc\u0645 \u06a9\u0631\u062f.  \u062a\u0627 \u0627\u0648\u0646 \u0645\u0648\u0642\u0639 \u06a9\u062f\u0646\u0648\u06cc\u0633\u06cc \u0645\u0628\u0627\u0631\u06a9 \ud83d\ude42<\/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 12:02:04<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;16075&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;\u062d\u0644 \u0645\u0633\u0627\u0626\u0644 \u062a\u0648\u0627\u0644\u06cc \u0628\u0627 LSTM \u062f\u0631 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>\u062f\u0631 \u0627\u06cc\u0646 \u0645\u0642\u0627\u0644\u0647\u060c \u0631\u0648\u0634 \u0627\u0646\u062c\u0627\u0645 \u067e\u06cc\u0634 \u0628\u06cc\u0646\u06cc \u0633\u0631\u06cc \u0647\u0627\u06cc \u0632\u0645\u0627\u0646\u06cc \u0631\u0627 \u06a9\u0647 \u0628\u0631\u0627\u06cc \u062d\u0644 \u0645\u0633\u0627\u0626\u0644 \u062a\u0648\u0627\u0644\u06cc \u0627\u0633\u062a\u0641\u0627\u062f\u0647 \u0645\u06cc \u0634\u0648\u062f\u060c \u06cc\u0627\u062f \u062e\u0648\u0627\u0647\u06cc\u062f \u06af\u0631\u0641\u062a. \u067e\u06cc\u0634\u200c\u0628\u06cc\u0646\u06cc \u0633\u0631\u06cc\u200c\u0647\u0627\u06cc \u0632\u0645\u0627\u0646\u06cc \u0628\u0647 \u0646\u0648\u0639 \u0645\u0634\u06a9\u0644\u0627\u062a\u06cc \u0627\u0634\u0627\u0631\u0647 \u0645\u06cc\u200c\u06a9\u0646\u062f \u06a9\u0647 \u062f\u0631 \u0622\u0646\u0647\u0627 \u0628\u0627\u06cc\u062f \u06cc\u06a9 \u0646\u062a\u06cc\u062c\u0647 \u0631\u0627 \u067e\u06cc\u0634\u200c\u0628\u06cc\u0646\u06cc \u06a9\u0646\u06cc\u0645 \u0631\u0648\u06cc \u0648\u0631\u0648\u062f\u06cc \u0647\u0627\u06cc \u0648\u0627\u0628\u0633\u062a\u0647 \u0628\u0647 \u0632\u0645\u0627\u0646 \u06cc\u06a9 \u0645\u062b\u0627\u0644 \u0645\u0639\u0645\u0648\u0644\u06cc \u0627\u0632 \u062f\u0627\u062f\u0647 \u0647\u0627\u06cc \u0633\u0631\u06cc \u0632\u0645\u0627\u0646\u06cc\u060c \u062f\u0627\u062f\u0647 \u0647\u0627\u06cc [&hellip;]<\/p>\n","protected":false},"author":3,"featured_media":9759,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[1743,620],"tags":[],"class_list":["post-16075","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\/16075","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=16075"}],"version-history":[{"count":0,"href":"https:\/\/rasanegaar.com\/blog\/wp-json\/wp\/v2\/posts\/16075\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/rasanegaar.com\/blog\/wp-json\/wp\/v2\/media\/9759"}],"wp:attachment":[{"href":"https:\/\/rasanegaar.com\/blog\/wp-json\/wp\/v2\/media?parent=16075"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/rasanegaar.com\/blog\/wp-json\/wp\/v2\/categories?post=16075"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/rasanegaar.com\/blog\/wp-json\/wp\/v2\/tags?post=16075"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}