\uc791\uc131\uc790: \uc870\uad11\ud615 | \ub0a0\uc9dc: [\ub0a0\uc9dc]<\/p>\n<\/header>\n \n \ucd5c\uadfc \uba87 \ub144\uac04 \ub525 \ub7ec\ub2dd \uae30\uc220\uc774 \ube44\uc57d\uc801\uc73c\ub85c \ubc1c\uc804\ud558\uba74\uc11c \ub2e4\uc591\ud55c \ubd84\uc57c\uc5d0 \ud65c\uc6a9\ub418\uace0 \uc788\uc2b5\ub2c8\ub2e4. \n \uc790\uc5f0\uc5b4 \ucc98\ub9ac(NLP)\ub294 \uc778\uacf5\uc9c0\ub2a5(AI)\uc758 \ud55c \ubd84\uc57c\ub85c, \uae30\uacc4\uac00 \uc790\uc5f0\uc5b4\ub97c \uc774\ud574\ud558\uace0 \n \uc2a4\ud338 \uba54\uc77c \ubd84\ub958\ub294 \ud2b9\ud788 \uc911\uc694\ud55c NLP \uc791\uc5c5 \uc911 \ud558\ub098\ub85c, \n 1D CNN\uc740 \uc8fc\ub85c \uc2dc\ud000\uc2a4 \ub370\uc774\ud130\uc5d0 \uc801\uc6a9\ub418\ub294 \uc2e0\uacbd\ub9dd \uad6c\uc870\ub85c, \ud14d\uc2a4\ud2b8 \ub370\uc774\ud130\uc640 \uac19\uc740 1\ucc28\uc6d0 \ub370\uc774\ud130\ub97c \ucc98\ub9ac\ud558\ub294 \ub370 \ud6a8\uacfc\uc801\uc785\ub2c8\ub2e4. \n 1D CNN\uc744 \uc0ac\uc6a9\ud558\uba74 \ud14d\uc2a4\ud2b8 \ub0b4\uc758 \uc9c0\uc5ed\uc801\uc778 \ud328\ud134\uc744 \ud6a8\uc728\uc801\uc73c\ub85c \ud559\uc2b5\ud560 \uc218 \uc788\uc2b5\ub2c8\ub2e4. \ub530\ub77c\uc11c \uc2a4\ud338 \uba54\uc77c \ubd84\ub958\uc640 \uac19\uc740 \uc790\uc5f0\uc5b4 \ucc98\ub9ac \uc791\uc5c5\uc5d0 \uc801\ud569\ud569\ub2c8\ub2e4.\n <\/p>\n<\/section>\n \n \uc2a4\ud338 \uba54\uc77c \ubd84\ub958\ub97c \uc704\ud574 \uc0ac\uc6a9\ud560 \uc218 \uc788\ub294 \ub2e4\uc591\ud55c \ub370\uc774\ud130\uc14b\uc774 \uc788\uc2b5\ub2c8\ub2e4. \n \ub370\uc774\ud130\uc14b\uc744 \uc900\ube44\ud558\uae30 \uc704\ud574\uc11c\ub294 \uba3c\uc800 \ub370\uc774\ud130\ub97c \uc218\uc9d1\ud558\uace0, \n \uc2a4\ud338 \uba54\uc77c \ubd84\ub958 \ubaa8\ub378\uc744 \uad6c\ucd95\ud558\uae30 \uc704\ud55c \uccab \ubc88\uc9f8 \ub2e8\uacc4\ub294 \ud14d\uc2a4\ud2b8 \ub370\uc774\ud130\ub97c \uc804\ucc98\ub9ac\ud558\ub294 \uac83\uc785\ub2c8\ub2e4. \n \uc774\ub7ec\ud55c \uc804\ucc98\ub9ac \uacfc\uc815\uc744 \uac70\uce5c \ud6c4\uc5d0\ub294 \uac01 \ub2e8\uc5b4\ub97c \ubca1\ud130\ub85c \ubcc0\ud658\ud574\uc57c \ud569\ub2c8\ub2e4. \n \uc774\uc81c, \uc804\ucc98\ub9ac\ub41c \ub370\uc774\ud130\ub97c \uae30\ubc18\uc73c\ub85c 1D CNN \ubaa8\ub378\uc744 \uc124\uacc4\ud558\uace0 \ud559\uc2b5\ud560 \ucc28\ub840\uc785\ub2c8\ub2e4. \n 1D CNN \ubaa8\ub378\uc740 \ucc28\ub840\ub85c \ud569\uc131\uacf1 \uce35, \ud480\ub9c1 \uce35, \uc644\uc804 \uc5f0\uacb0 \uce35\uc73c\ub85c \uad6c\uc131\ub429\ub2c8\ub2e4. \n \uc704\uc758 \ucf54\ub4dc\uc5d0\uc11c \uc784\ubca0\ub529 \uce35\uc744 \ud1b5\ud574 \ub2e8\uc5b4 \uc784\ubca0\ub529\uc744 \uc218\ud589\ud558\uace0, \n \ubaa8\ub378\uc744 \ucef4\ud30c\uc77c\ud558\uace0 \ud559\uc2b5\ud558\uae30 \uc704\ud574\uc11c\ub294 \uc190\uc2e4 \ud568\uc218\uc640 \ucd5c\uc801\ud654 \uc54c\uace0\ub9ac\uc998\uc744 \uc124\uc815\ud574\uc57c \ud569\ub2c8\ub2e4. \n \ud559\uc2b5\ub41c \ubaa8\ub378\uc740 \ud14c\uc2a4\ud2b8 \ub370\uc774\ud130\uc14b\uc744 \uc0ac\uc6a9\ud558\uc5ec \uc131\ub2a5\uc744 \ud3c9\uac00\ud560 \uc218 \uc788\uc2b5\ub2c8\ub2e4. \n \ubaa8\ub378\uc758 \uc131\ub2a5\uc744 \ud3c9\uac00\ud558\uae30 \uc704\ud574\uc11c\ub294 \ud14c\uc2a4\ud2b8 \ub370\uc774\ud130\uc14b\uc744 \ud65c\uc6a9\ud569\ub2c8\ub2e4. \n \ub2e4\uc74c \ucf54\ub4dc\ub294 \ubaa8\ub378\uc758 \uc131\ub2a5\uc744 \ud3c9\uac00\ud558\ub294 \ubc29\ubc95\uc744 \ubcf4\uc5ec\uc90d\ub2c8\ub2e4:\n <\/p>\n \n \uc774\ub97c \ud1b5\ud574 \ubaa8\ub378\uc774 \uc5bc\ub9c8\ub098 \uc798 \ubd84\ub958\ud558\ub294\uc9c0 \uad6c\uccb4\uc801\uc778 \uc218\uce58\ub97c \ud655\uc778\ud560 \uc218 \uc788\uc2b5\ub2c8\ub2e4.\n <\/p>\n<\/section>\n \n \ubcf8 \uae00\uc5d0\uc11c\ub294 1D CNN\uc744 \ud65c\uc6a9\ud558\uc5ec \uc2a4\ud338 \uba54\uc77c\uc744 \ubd84\ub958\ud558\ub294 \ubc29\ubc95\uc5d0 \ub300\ud574 \uc54c\uc544\ubcf4\uc558\uc2b5\ub2c8\ub2e4.\uc11c\ub860<\/h2>\n
\n \uadf8 \uc911\uc5d0\uc11c\ub3c4 \uc790\uc5f0\uc5b4 \ucc98\ub9ac(Natural Language Processing, NLP)\ub294 \ucef4\ud4e8\ud130\uac00 \uc778\uac04\uc758 \uc5b8\uc5b4\ub97c \uc774\ud574\ud558\uace0 \uc0dd\uc131\ud558\ub294 \uae30\uc220\ub85c,
\n \uc774\uba54\uc77c \ubd84\ub958, \uac10\uc815 \ubd84\uc11d, \uae30\uacc4 \ubc88\uc5ed \ub4f1 \ub2e4\ubc29\uba74\uc5d0\uc11c \ud65c\uc6a9\ub418\uace0 \uc788\uc2b5\ub2c8\ub2e4.
\n \ubcf8 \uae00\uc5d0\uc11c\ub294 1\ucc28\uc6d0 \ud569\uc131\uacf1 \uc2e0\uacbd\ub9dd(1D CNN)\uc744 \uc0ac\uc6a9\ud558\uc5ec \uc2a4\ud338 \uba54\uc77c\uc744 \ubd84\ub958\ud558\ub294 \ubc29\ubc95\uc5d0 \ub300\ud574
\n \uc790\uc138\ud788 \uc124\uba85\ud558\uace0\uc790 \ud569\ub2c8\ub2e4. \uc6b0\ub9ac\ub294 \uba3c\uc800 \uc790\uc5f0\uc5b4 \ucc98\ub9ac\uc758 \uae30\ucd08\ub97c \uc0b4\ud3b4\ubcf8 \ub4a4,
\n 1D CNN\uc758 \uad6c\uc870\uc640 \uc801\uc6a9 \ubc29\ubc95\uc744 \uc774\ud574\ud558\uace0, \ub9c8\uc9c0\ub9c9\uc73c\ub85c \uc2e4\uc2b5\uc744 \ud1b5\ud574 \uc2a4\ud338 \uba54\uc77c \ubd84\ub958\uae30\ub97c \uad6c\ucd95\ud574\ubcf4\uaca0\uc2b5\ub2c8\ub2e4.\n <\/p>\n<\/section>\n\uc790\uc5f0\uc5b4 \ucc98\ub9ac(NLP)\ub780?<\/h2>\n
\n \ud574\uc11d\ud558\ub3c4\ub85d \ub3d5\ub294 \uae30\uc220\uc785\ub2c8\ub2e4. \uc790\uc5f0\uc5b4 \ucc98\ub9ac\uc758 \uc8fc\uc694 \uc791\uc5c5\uc5d0\ub294 \ub2e4\uc74c\uacfc \uac19\uc740 \uac83\ub4e4\uc774 \uc788\uc2b5\ub2c8\ub2e4:\n <\/p>\n\n
\n \uc0ac\uc6a9\uc790\uc5d0\uac8c \uc6d0\uce58 \uc54a\ub294 \uc774\uba54\uc77c\uc744 \ud544\ud130\ub9c1\ud558\uc5ec \ud6a8\uc728\uc801\uc778 \uc774\uba54\uc77c \uad00\ub9ac\ub97c \uac00\ub2a5\ud558\uac8c \ud569\ub2c8\ub2e4.
\n \uc804\ud1b5\uc801\uc73c\ub85c \uc774\ub7ec\ud55c \ubd84\ub958 \uc791\uc5c5\uc740 \uaddc\uce59 \uae30\ubc18 \uc811\uadfc \ub610\ub294 \uba38\uc2e0 \ub7ec\ub2dd \uae30\ubc95\uc744 \uc0ac\uc6a9\ud558\uc5ec \uc218\ud589\ub418\uc5b4 \uc654\uc73c\ub098,
\n \ucd5c\uadfc\uc5d0\ub294 \ub525 \ub7ec\ub2dd \uae30\uc220\uc774 \uc774\ub7ec\ud55c \ubb38\uc81c\ub97c \ud574\uacb0\ud558\ub294 \ub370 \uc788\uc5b4 \ub192\uc740 \uc131\ub2a5\uc744 \ubcf4\uc5ec\uc8fc\uace0 \uc788\uc2b5\ub2c8\ub2e4.\n <\/p>\n<\/section>\n1D CNN(1\ucc28\uc6d0 \ud569\uc131\uacf1 \uc2e0\uacbd\ub9dd) \uc18c\uac1c<\/h2>\n
\n CNN\uc740 \uc8fc\ub85c \uc774\ubbf8\uc9c0 \uc778\uc2dd\uc5d0 \uc0ac\uc6a9\ub418\uc9c0\ub9cc, \uc2dc\ud000\uc2a4 \ub370\uc774\ud130\uc5d0\ub3c4 \uc751\uc6a9 \uac00\ub2a5\ud569\ub2c8\ub2e4. 1D CNN\uc758 \uc8fc\uc694 \uad6c\uc131 \uc694\uc18c\ub294 \ub2e4\uc74c\uacfc \uac19\uc2b5\ub2c8\ub2e4:\n <\/p>\n\n
\uc2a4\ud338 \uba54\uc77c \ubd84\ub958\uc758 \ub370\uc774\ud130\uc14b \uc900\ube44<\/h2>\n
\n \uc608\ub97c \ub4e4\uc5b4, SMS Spam Collection \ub370\uc774\ud130\uc14b<\/a>\ub97c
\n \uc0ac\uc6a9\ud560 \uc218 \uc788\uc73c\uba70, \uc774\uba54\uc77c \ub370\uc774\ud130\uc14b\uc73c\ub85c\ub294 Spambase \ub370\uc774\ud130\uc14b<\/a>\uac00 \uc788\uc2b5\ub2c8\ub2e4.
\n \uc774\ub7ec\ud55c \ub370\uc774\ud130\uc14b\uc5d0\ub294 \uc2a4\ud338 \ubc0f \ube44\uc2a4\ud338\uc73c\ub85c \ub77c\ubca8\ub9c1\ub41c \uc774\uba54\uc77c \ub610\ub294 \uba54\uc2dc\uc9c0\uac00 \ud3ec\ud568\ub418\uc5b4 \uc788\uc2b5\ub2c8\ub2e4.\n <\/p>\n
\n \ub370\uc774\ud130 \ud074\ub9ac\ub2dd(Data Cleaning)\uacfc \uc804\ucc98\ub9ac(Preprocessing) \uacfc\uc815\uc744 \uac70\uccd0\uc57c \ud569\ub2c8\ub2e4.
\n \uc774 \uacfc\uc815\uc5d0\ub294 \ud2b9\uc218 \ubb38\uc790 \ubc0f \ubd88\uc6a9\uc5b4(stop words) \uc81c\uac70,
\n \ud14d\uc2a4\ud2b8 \uc18c\ubb38\uc790\ud654, \ud1a0\ud070\ud654(tokenization) \uacfc\uc815\uc774 \ud3ec\ud568\ub429\ub2c8\ub2e4.\n <\/p>\n<\/section>\n\ud14d\uc2a4\ud2b8 \uc804\ucc98\ub9ac \uacfc\uc815<\/h2>\n
\n \uc804\ucc98\ub9ac \uc808\ucc28\ub294 \ub2e4\uc74c\uacfc \uac19\uc740 \ub2e8\uacc4\ub85c \uc774\ub8e8\uc5b4\uc9d1\ub2c8\ub2e4:\n <\/p>\n\n
\n \uc77c\ubc18\uc801\uc73c\ub85c \uc0ac\uc6a9\ub418\ub294 \ubc29\ubc95\uc740 \ub2e8\uc5b4 \uc784\ubca0\ub529(Word Embedding)<\/strong> \uae30\ubc95\uc73c\ub85c,
\n \ub300\ud45c\uc801\uc73c\ub85c\ub294 Word2Vec, GloVe, FastText \ubaa8\ub378\uc774 \uc788\uc2b5\ub2c8\ub2e4.
\n \uc774\ub97c \ud1b5\ud574 \ub2e8\uc5b4\ub294 \uace0\ucc28\uc6d0 \uacf5\uac04\uc5d0\uc11c\uc758 \ubca1\ud130\ub85c \ud45c\ud604\ub418\uba70, \uc720\uc0ac\ud55c \uc758\ubbf8\ub97c \uac00\uc9c4 \ub2e8\uc5b4\ub07c\ub9ac \uac00\uae5d\uac8c \ubc30\uce58\ub429\ub2c8\ub2e4.\n <\/p>\n<\/section>\n\ubaa8\ub378 \uc124\uacc4 \ubc0f \ud559\uc2b5<\/h2>\n
\n Keras\uc640 TensorFlow\ub97c \ud65c\uc6a9\ud558\uc5ec \uc2a4\ud338 \uba54\uc77c \ubd84\ub958 \ubaa8\ub378\uc744 \uad6c\ucd95\ud558\ub294 \ubc29\ubc95\uc740 \ub2e4\uc74c\uacfc \uac19\uc2b5\ub2c8\ub2e4:\n <\/p>\n1. \ubaa8\ub378 \uc124\uacc4<\/h3>\n
\n \uc608\uc2dc \ucf54\ub4dc\ub85c \ubaa8\ub378\uc758 \uad6c\uc870\ub97c \ub2e4\uc74c\uacfc \uac19\uc774 \uc815\uc758\ud560 \uc218 \uc788\uc2b5\ub2c8\ub2e4:\n <\/p>\n\n
\nfrom keras.models import Sequential\nfrom keras.layers import Conv1D, MaxPooling1D, Flatten, Dense, Embedding, Dropout\n\nmodel = Sequential()\nmodel.add(Embedding(input_dim=vocab_size, output_dim=embedding_dim, input_length=max_length))\nmodel.add(Conv1D(filters=64, kernel_size=5, activation='relu'))\nmodel.add(MaxPooling1D(pool_size=2))\nmodel.add(Flatten())\nmodel.add(Dense(10, activation='relu'))\nmodel.add(Dropout(0.5))\nmodel.add(Dense(1, activation='sigmoid'))\n <\/code>\n <\/pre>\n
\n \ud569\uc131\uacf1 \uce35\uc744 \ud1b5\ud574 \ud2b9\uc9d5\uc744 \ucd94\ucd9c\ud55c \ud6c4 \ud480\ub9c1 \uce35\uc5d0\uc11c \ucc28\uc6d0\uc744 \ucd95\uc18c\ud569\ub2c8\ub2e4.
\n \ub9c8\uc9c0\ub9c9\uc73c\ub85c \ucd9c\ub825\uce35\uc5d0\uc11c \uc2a4\ud338\/\ube44\uc2a4\ud338 \uc5ec\ubd80\ub97c \ubd84\ub958\ud569\ub2c8\ub2e4.\n <\/p>\n2. \ubaa8\ub378 \ucef4\ud30c\uc77c \ubc0f \ud559\uc2b5<\/h3>\n
\n \uc77c\ubc18\uc801\uc73c\ub85c \uc774\uc9c4 \ubd84\ub958\uc758 \uacbd\uc6b0 binary_crossentropy \uc190\uc2e4 \ud568\uc218\ub97c \uc0ac\uc6a9\ud569\ub2c8\ub2e4.
\n \ub2e4\uc74c \ucf54\ub4dc\ub294 \ubaa8\ub378\uc744 \ucef4\ud30c\uc77c\ud558\uace0 \ud559\uc2b5\ud558\ub294 \ubc29\ubc95\uc744 \ubcf4\uc5ec\uc90d\ub2c8\ub2e4:\n <\/p>\n\n
\nmodel.compile(optimizer='adam', loss='binary_crossentropy', metrics=['accuracy'])\nmodel.fit(X_train, y_train, epochs=10, batch_size=32, validation_split=0.2)\n <\/code>\n <\/pre>\n
\n \ud3c9\uac00 \uacb0\uacfc\ub294 \uc815\ud655\ub3c4 \ubc0f \uc190\uc2e4 \uac12\uc73c\ub85c \ud655\uc778\ud560 \uc218 \uc788\uc2b5\ub2c8\ub2e4.\n <\/p>\n<\/section>\n\ubaa8\ub378 \uc131\ub2a5 \ud3c9\uac00<\/h2>\n
\n \uc77c\ubc18\uc801\uc73c\ub85c F1 Score, Precision, Recall\uacfc \uac19\uc740 \uc9c0\ud45c\ub97c \uc0ac\uc6a9\ud558\uc5ec \ubaa8\ub378\uc744 \ud3c9\uac00\ud569\ub2c8\ub2e4.\n <\/p>\n1. \ud3c9\uac00 \uc9c0\ud45c \uc124\uba85<\/h3>\n
\n
2. \uc131\ub2a5 \ud3c9\uac00 \ucf54\ub4dc<\/h3>\n
\n
\nfrom sklearn.metrics import classification_report\n\ny_pred = model.predict(X_test)\ny_pred_classes = (y_pred > 0.5).astype(\"int32\")\nprint(classification_report(y_test, y_pred_classes))\n <\/code>\n <\/pre>\n\uacb0\ub860<\/h2>\n
\n \ub525 \ub7ec\ub2dd \ubc0f CNN \uad6c\uc870\uc758 \uc774\ud574\uc640 \ud568\uaed8 \uc790\uc5f0\uc5b4 \ucc98\ub9ac\uc758 \uae30\ucd08 \uae30\uc220\uc744 \uc801\uc6a9\ud558\uc5ec
\n \uc2a4\ud338 \uba54\uc77c \ubd84\ub958\uae30\ub97c \uad6c\ucd95\ud558\uace0 \ud3c9\uac00\ud558\ub294 \uacfc\uc815\uc744 \uc124\uba85\ud588\uc2b5\ub2c8\ub2e4.
\n \uc774\ub7ec\ud55c \uae30\uc220\uc740 \ud5a5\ud6c4 \ub354 \ubcf5\uc7a1\ud55c \uc790\uc5f0\uc5b4 \ucc98\ub9ac \ubb38\uc81c\ub97c \ud574\uacb0\ud558\ub294 \ub370 \uc720\uc6a9\ud560 \uac83\uc785\ub2c8\ub2e4.
\n \uc55e\uc73c\ub85c\ub3c4 \ub525 \ub7ec\ub2dd\uc774 \uc778\uacf5\uc9c0\ub2a5 \ubd84\uc57c\uc5d0\uc11c \uc5b4\ub5a4 \ud601\uc2e0\uc744 \uac00\uc838\uc62c\uc9c0 \uae30\ub300\ub429\ub2c8\ub2e4.\n <\/p>\n<\/section>\n