{"id":28656,"date":"2024-10-27T16:14:22","date_gmt":"2024-10-27T16:14:22","guid":{"rendered":"http:\/\/atmokpo.com\/w\/?p=28656"},"modified":"2024-11-26T06:59:39","modified_gmt":"2024-11-26T06:59:39","slug":"%eb%a8%b8%ec%8b%a0%eb%9f%ac%eb%8b%9d-%eb%b0%8f-%eb%94%a5%eb%9f%ac%eb%8b%9d-%ec%95%8c%ea%b3%a0%eb%a6%ac%ec%a6%98-%ed%8a%b8%eb%a0%88%ec%9d%b4%eb%94%a9-doc2vec-%eb%aa%a8%eb%8d%b8-%ed%9b%88%eb%a0%a8","status":"publish","type":"post","link":"https:\/\/atmokpo.com\/w\/28656\/","title":{"rendered":"\uba38\uc2e0\ub7ec\ub2dd \ubc0f \ub525\ub7ec\ub2dd \uc54c\uace0\ub9ac\uc998 \ud2b8\ub808\uc774\ub529, doc2vec \ubaa8\ub378 \ud6c8\ub828"},"content":{"rendered":"

<\/p>\n

\uc624\ub298\ub0a0 \uae08\uc735 \uc2dc\uc7a5\uc5d0\uc11c\uc758 \ud2b8\ub808\uc774\ub529 \uc804\ub7b5\uc740 \ub370\uc774\ud130\uc758 \uc591\uacfc \ubcf5\uc7a1\uc131\uc73c\ub85c \uc778\ud574 \ub354\uc6b1 \uc5b4\ub824\uc6cc\uc9c0\uace0 \uc788\uc2b5\ub2c8\ub2e4. \uc774\uc640 \uac19\uc740 \ud658\uacbd\uc5d0\uc11c \uba38\uc2e0\ub7ec\ub2dd\uacfc \ub525\ub7ec\ub2dd \uc54c\uace0\ub9ac\uc998\uc740 \ud2b8\ub808\uc774\ub529\uc5d0\uc11c \uc911\uc694\ud55c \ub3c4\uad6c\ub85c \uc790\ub9ac \uc7a1\uace0 \uc788\uc2b5\ub2c8\ub2e4. \ubcf8 \uac15\uc88c\uc5d0\uc11c\ub294 \uc790\uc5f0\uc5b4 \ucc98\ub9ac \uae30\uc220 \uc911 \ud558\ub098\uc778 doc2vec<\/strong> \ubaa8\ub378\uc744 \uc0ac\uc6a9\ud558\uc5ec \ud14d\uc2a4\ud2b8 \ub370\uc774\ud130\ub97c \ubca1\ud130\ub85c \ubcc0\ud658\ud558\uace0 \uc774\ub97c \ud1b5\ud574 \ud2b8\ub808\uc774\ub529 \uc2e0\ud638\ub97c \uc0dd\uc131\ud558\ub294 \ubc29\ubc95\uc744 \uc54c\uc544\ubcf4\uaca0\uc2b5\ub2c8\ub2e4.<\/p>\n

1. \uc778\uacf5\uc9c0\ub2a5 \uae30\ubc18 \ud2b8\ub808\uc774\ub529\uc758 \uae30\ucd08<\/h2>\n

\uc778\uacf5\uc9c0\ub2a5 \uae30\ubc18 \ud2b8\ub808\uc774\ub529\uc758 \uae30\ubcf8 \uac1c\ub150\uc740 \ub370\uc774\ud130\uc5d0\uc11c \ud328\ud134\uc744 \ucc3e\uc544\ub0b4\uc5b4 \uc774\ub97c \ud2b8\ub808\uc774\ub529 \uc2e0\ud638\ub85c \ubcc0\ud658\ud558\ub294 \uac83\uc785\ub2c8\ub2e4. \uacfc\uac70\uc758 \uac00\uaca9 \ub370\uc774\ud130, \ub274\uc2a4, \uc18c\uc15c\ubbf8\ub514\uc5b4 \ub4f1\uc758 \ub2e4\uc591\ud55c \ub370\uc774\ud130\ub97c \ud65c\uc6a9\ud558\uc5ec \uc54c\uace0\ub9ac\uc998\uc744 \ud1b5\ud574 \uc758\uc0ac \uacb0\uc815\uc744 \ub0b4\ub9bd\ub2c8\ub2e4. \uc774 \uacfc\uc815\uc5d0\uc11c \uba38\uc2e0\ub7ec\ub2dd\uacfc \ub525\ub7ec\ub2dd \uae30\uc220\uc774 \uc8fc\ub85c \uc0ac\uc6a9\ub429\ub2c8\ub2e4.<\/p>\n

2. doc2vec \ubaa8\ub378 \uc774\ud574\ud558\uae30<\/h2>\n

doc2vec\uc740 \ub2e8\uc5b4 \ubca1\ud130 \ubaa8\ub378\uc758 \ud655\uc7a5\uc73c\ub85c, \ubb38\uc11c \uc804\uccb4\ub97c \ud558\ub098\uc758 \ubca1\ud130\ub85c \ud45c\ud604\ud560 \uc218 \uc788\ub294 \ubc29\ubc95\uc785\ub2c8\ub2e4. \uc774\ub294 \ub300\ub7c9\uc758 \ud14d\uc2a4\ud2b8 \ub370\uc774\ud130\ub97c \ubcf4\ub2e4 \ud6a8\uc728\uc801\uc73c\ub85c \ucc98\ub9ac\ud558\uace0, \ubb38\uc11c \uac04\uc758 \uc720\uc0ac\uc131\uc744 \uacc4\uc0b0\ud558\ub294 \ub370 \uc720\uc6a9\ud569\ub2c8\ub2e4. Gensim<\/strong> \ub77c\uc774\ube0c\ub7ec\ub9ac\ub97c \uc0ac\uc6a9\ud558\uc5ec doc2vec \ubaa8\ub378\uc744 \uad6c\ucd95\ud558\uace0 \ud6c8\ub828\ud560 \uc218 \uc788\uc2b5\ub2c8\ub2e4.<\/p>\n

2.1 doc2vec\uc758 \uc6d0\ub9ac<\/h3>\n

doc2vec\uc740 \ub450 \uac00\uc9c0 \uc8fc\uc694 \uc811\uadfc\ubc95\uc778 Distributed Bag of Words (DBOW)<\/code>\uc640 Distributed Memory (DM)<\/code>\ub97c \uc0ac\uc6a9\ud558\uc5ec \ubb38\uc11c \uc784\ubca0\ub529\uc744 \uc0dd\uc131\ud569\ub2c8\ub2e4. DBOW\ub294 \uc8fc\uc5b4\uc9c4 \ubb38\uc11c\uc5d0\uc11c \ud2b9\uc815 \ub2e8\uc5b4\ub4e4\uc744 \uc608\uce21\ud558\ub294 \ubaa8\ub378\uc774\uba70, DM\uc740 \uc8fc\uc5b4\uc9c4 \ub2e8\uc5b4\ub85c\ubd80\ud130 \ubb38\uc11c\ub97c \uc608\uce21\ud558\ub294 \ubc29\ubc95\uc785\ub2c8\ub2e4. \uc774\ub7ec\ud55c \ubaa8\ub378\uc758 \ud559\uc2b5\uc744 \ud1b5\ud574 \uac01 \ubb38\uc11c\ub294 \uace0\ucc28\uc6d0 \ubca1\ud130\ub85c \ubcc0\ud658\ub429\ub2c8\ub2e4.<\/p>\n

2.2 doc2vec \uad6c\ud604\ud558\uae30<\/h3>\n
import gensim\nfrom gensim.models.doc2vec import Doc2Vec, TaggedDocument\n\n# \ubb38\uc11c \ub370\uc774\ud130\ub97c \uc900\ube44\ud569\ub2c8\ub2e4.\ndocuments = [\n    TaggedDocument(words=['\uc774', '\uac83\uc740', '\uccab\ubc88\uc9f8', '\ubb38\uc11c', '\uc785\ub2c8\ub2e4.'], tags=['doc1']),\n    TaggedDocument(words=['\ub450\ubc88\uc9f8', '\ubb38\uc11c\uac00', '\uc5ec\uae30', '\uc788\uc2b5\ub2c8\ub2e4.'], tags=['doc2']),\n    TaggedDocument(words=['\uc138\ubc88\uc9f8', '\ubb38\uc11c\uc785\ub2c8\ub2e4.'], tags=['doc3'])\n]\n\n# doc2vec \ubaa8\ub378 \uc0dd\uc131\nmodel = Doc2Vec(vector_size=20, min_count=2, epochs=100)\n\n# \ubaa8\ub378\uc5d0 \ubb38\uc11c \ucd94\uac00\nmodel.build_vocab(documents)\nmodel.train(documents, total_examples=model.corpus_count, epochs=model.epochs)\n<\/code><\/pre>\n
3. \ub370\uc774\ud130 \uc900\ube44 \ubc0f \uc804\ucc98\ub9ac<\/h2>\n
doc2vec \ubaa8\ub378\uc744 \ud6c8\ub828\uc2dc\ud0a4\uae30 \uc704\ud574\uc11c\ub294 \ud488\uc9c8 \ub192\uc740 \ud14d\uc2a4\ud2b8 \ub370\uc774\ud130\uac00 \ud544\uc694\ud569\ub2c8\ub2e4. \uc8fc\uc2dd \uc2dc\uc7a5 \ub370\uc774\ud130, \ub274\uc2a4 \uae30\uc0ac, \uc18c\uc15c\ubbf8\ub514\uc5b4 \ud3ec\uc2a4\ud2b8 \ub4f1\uc744 \uc218\uc9d1\ud558\uace0 \uc774\ub97c \uc804\ucc98\ub9ac\ud574\uc57c \ud569\ub2c8\ub2e4. \uc804\ucc98\ub9ac \uacfc\uc815\uc5d0\uc11c\ub294 \ubd88\uc6a9\uc5b4 \uc81c\uac70, \ud1a0\ud070\ud654, \ud45c\uc81c\uc5b4 \ucd94\ucd9c \ub4f1\uc774 \ud3ec\ud568\ub429\ub2c8\ub2e4.<\/p>\n
3.1 \ud14d\uc2a4\ud2b8 \ub370\uc774\ud130 \uc218\uc9d1<\/h3>\n
\uc5ec\ub7ec \uc18c\uc2a4\uc5d0\uc11c \ub370\uc774\ud130\ub97c \uc218\uc9d1\ud560 \uc218 \uc788\uc2b5\ub2c8\ub2e4. \uc608\ub97c \ub4e4\uc5b4, Yahoo Finance API\ub098 Twitter API\ub97c \ud65c\uc6a9\ud558\uc5ec \uc2e4\uc2dc\uac04 \ub274\uc2a4\uc640 \ud2b8\uc704\ud130 \ub370\uc774\ud130\ub97c \uc218\uc9d1\ud560 \uc218 \uc788\uc2b5\ub2c8\ub2e4.<\/p>\n
3.2 \ub370\uc774\ud130 \uc804\ucc98\ub9ac<\/h3>\n
import nltk\nfrom nltk.corpus import stopwords\nfrom nltk.tokenize import word_tokenize\n\nnltk.download('punkt')\nnltk.download('stopwords')\nstop_words = set(stopwords.words('korean'))\n\ndef preprocess_text(text):\n    # \ud1a0\ud070\ud654\n    words = word_tokenize(text)\n    # \ubd88\uc6a9\uc5b4 \uc81c\uac70\n    filtered_words = [word for word in words if word not in stop_words]\n    return filtered_words\n\n# \uc608\uc81c \ubb38\uc11c \uc804\ucc98\ub9ac\ntexts = [\"\uc5ec\uae30\uc5d0\ub294 \uc8fc\uc2dd \uc2dc\uc7a5\uc758 \uac00\uc7a5 \ucd5c\uadfc \ub274\uc2a4\uac00 \ud3ec\ud568\ub418\uc5b4 \uc788\uc2b5\ub2c8\ub2e4.\"]\nprocessed_texts = [preprocess_text(text) for text in texts]\n<\/code><\/pre>\n
4. \ubaa8\ub378 \ud6c8\ub828 \ubc0f \ud3c9\uac00<\/h2>\n
\ubaa8\ub378\uc744 \ud6c8\ub828\uc2dc\ud0a8 \ud6c4, \uc6b0\ub9ac\ub294 \ubaa8\ub378\uc758 \uc131\ub2a5\uc744 \ud3c9\uac00\ud558\uace0 \uc801\uc808\ud558\uac8c \ud29c\ub2dd\ud574\uc57c \ud569\ub2c8\ub2e4. \uac00\uc7a5 \uc77c\ubc18\uc801\uc778 \ud3c9\uac00\uc9c0\ud45c\ub294 \ubb38\uc11c\uc758 \uc720\uc0ac\ub3c4\ub97c \uce21\uc815\ud558\uc5ec \ud655\uc778\ud558\ub294 \uac83\uc785\ub2c8\ub2e4.<\/p>\n
4.1 \ubaa8\ub378 \ud6c8\ub828<\/h3>\n
model.train(filtered_documents, total_examples=model.corpus_count, epochs=model.epochs)\n<\/code><\/pre>\n
4.2 \ubaa8\ub378 \ud3c9\uac00<\/h3>\n
\ud6c8\ub828\ub41c \ubaa8\ub378\uc744 \ud1b5\ud574 \uc0c8\ub85c\uc6b4 \ubb38\uc11c\uc758 \ubca1\ud130\ub97c \uc0dd\uc131\ud558\uace0, KNN \ub610\ub294 Cosine Similarity \ub4f1\uc758 \uae30\ubc95\uc744 \uc0ac\uc6a9\ud558\uc5ec \uc720\uc0ac\ub3c4\ub97c \ud3c9\uac00\ud560 \uc218 \uc788\uc2b5\ub2c8\ub2e4.<\/p>\n
5. \ud2b8\ub808\uc774\ub529 \uc2e0\ud638 \uc0dd\uc131<\/h2>\n
doc2vec\uc744 \ud1b5\ud574 \uc0dd\uc131\ub41c \ubb38\uc11c \ubca1\ud130\ub97c \uae30\ubc18\uc73c\ub85c \uba38\uc2e0\ub7ec\ub2dd \uc54c\uace0\ub9ac\uc998\uc744 \ud65c\uc6a9\ud558\uc5ec \ud2b8\ub808\uc774\ub529 \uc2e0\ud638\ub97c \uc0dd\uc131\ud560 \uc218 \uc788\uc2b5\ub2c8\ub2e4. \uc608\ub97c \ub4e4\uc5b4, \ubb38\uc11c\uc758 \uac10\uc131\uc744 \ubd84\uc11d\ud558\uc5ec \ub9e4\ub9e4 \ubc29\ud5a5\uc131\uc744 \uacb0\uc815\ud560 \uc218 \uc788\uc2b5\ub2c8\ub2e4.<\/p>\n
5.1 \uac10\uc131 \ubd84\uc11d \ubaa8\ub378 \uad6c\ucd95<\/h3>\n
\uac10\uc131 \ubd84\uc11d\uc744 \uc704\ud574 \ub79c\ub364 \ud3ec\ub808\uc2a4\ud2b8\ub098 SVM\uacfc \uac19\uc740 \uae30\uacc4\ub97c \uc0ac\uc6a9\ud560 \uc218 \uc788\uc73c\uba70, \uc774\ub97c \ud1b5\ud574 \uae0d\uc815\uc801, \ubd80\uc815\uc801 \uc2e0\ud638\ub97c \uad6c\ubd84\ud560 \uc218 \uc788\uc2b5\ub2c8\ub2e4.<\/p>\n
from sklearn.ensemble import RandomForestClassifier\n\n# \uac10\uc131 \ubd84\uc11d \ub370\uc774\ud130\uc14b \uad6c\uc131\nX = ...\ny = ...\n\n# \ub79c\ub364 \ud3ec\ub808\uc2a4\ud2b8 \ubaa8\ub378 \ud6c8\ub828\nrf_model = RandomForestClassifier(n_estimators=100)\nrf_model.fit(X, y)\n<\/code><\/pre>\n
5.2 \uc2e0\ud638 \uc0dd\uc131 \ubc0f \ud2b8\ub808\uc774\ub529 \uc804\ub7b5<\/h3>\n
\ud6c8\ub828\ub41c \uac10\uc131 \ubd84\uc11d \ubaa8\ub378\uc744 \uc0ac\uc6a9\ud558\uc5ec \uc2e4\uc2dc\uac04 \ub370\uc774\ud130\ub97c \ubd84\uc11d\ud558\uace0, \uc774\uc5d0 \ub530\ub978 \ud2b8\ub808\uc774\ub529 \uc2e0\ud638\ub97c \uc0dd\uc131\ud569\ub2c8\ub2e4. \uc774\ub97c \ud1b5\ud574 \uc790\ub3d9 \ub9e4\ub9e4 \uc2dc\uc2a4\ud15c\uc744 \uad6c\ucd95\ud560 \uc218 \uc788\uc2b5\ub2c8\ub2e4.<\/p>\n
6. \uc790\ub3d9\ub9e4\ub9e4 \uc2dc\uc2a4\ud15c \ud1b5\ud569<\/h2>\n
\ub9c8\uc9c0\ub9c9\uc73c\ub85c, \uc0dd\uc131\ub41c \ud2b8\ub808\uc774\ub529 \uc2e0\ud638\ub97c \ubc14\ud0d5\uc73c\ub85c \uc790\ub3d9\ub9e4\ub9e4 \uc2dc\uc2a4\ud15c\uc744 \ud1b5\ud569\ud574\uc57c \ud569\ub2c8\ub2e4. \ub2e4\uc591\ud55c \ud2b8\ub808\uc774\ub529 API\ub97c \ud65c\uc6a9\ud558\uc5ec \ub9e4\ub9e4\ub97c \uc2e4\ud589\ud560 \uc218 \uc788\uc2b5\ub2c8\ub2e4.<\/p>\n
import requests\n\ndef execute_trade(signal):\n    if signal == 'buy':\n        # \ub9e4\uc218 \uc8fc\ubb38 \uc2e4\ud589\n        requests.post(\"API_URL\/buy\", data= ...)\n    elif signal == 'sell':\n        # \ub9e4\ub3c4 \uc8fc\ubb38 \uc2e4\ud589\n        requests.post(\"API_URL\/sell\", data= ...)\n<\/code><\/pre>\n
7. \uacb0\ub860<\/h2>\n
\ubcf8 \uac15\uc88c\uc5d0\uc11c\ub294 \uba38\uc2e0\ub7ec\ub2dd\uacfc \ub525\ub7ec\ub2dd\uc744 \ud65c\uc6a9\ud55c doc2vec \ubaa8\ub378 \ud6c8\ub828 \ubc0f \ud14d\uc2a4\ud2b8 \ub370\uc774\ud130 \uae30\ubc18\uc758 \ud2b8\ub808\uc774\ub529 \uc2e0\ud638 \uc0dd\uc131\uc744 \uc0b4\ud3b4\ubcf4\uc558\uc2b5\ub2c8\ub2e4. \uc774 \uacfc\uc815\uc744 \ud1b5\ud574 \ub354\uc6b1 \uc815\uad50\ud55c \uc790\ub3d9\ub9e4\ub9e4 \uc804\ub7b5\uc744 \uad6c\ucd95\ud560 \uc218 \uc788\uc73c\uba70, \uc774\ub97c \ud1b5\ud574 \uae08\uc735 \uc2dc\uc7a5\uc5d0\uc11c\uc758 \uc131\uacfc\ub97c \uadf9\ub300\ud654\ud560 \uc218 \uc788\uc744 \uac83\uc785\ub2c8\ub2e4. \uc55e\uc73c\ub85c \ub354 \ubc1c\uc804\ud558\ub294 AI \uae30\uc220\uc744 \ud1b5\ud574 \uae08\uc735 \ubd84\uc57c\uc5d0\uc11c \uc0c8\ub85c\uc6b4 \uac00\ub2a5\uc131\uc744 \uc5f4\uc5b4\uac08 \uc218 \uc788\uae30\ub97c \uae30\ub300\ud569\ub2c8\ub2e4.<\/p>\n
\ucc38\uace0 \uc790\ub8cc<\/h2>\n