\ucd5c\uadfc \uae08\uc735 \uc2dc\uc7a5\uc5d0\uc11c \uba38\uc2e0\ub7ec\ub2dd\uacfc \ub525\ub7ec\ub2dd \uae30\uc220\uc744 \ud65c\uc6a9\ud55c \uc790\ub3d9\ub9e4\ub9e4(trading)\uac00 \uc8fc\ubaa9\ubc1b\uace0 \uc788\uc2b5\ub2c8\ub2e4. \ud22c\uc790\uc790\ub4e4\uc774 \ube60\ub974\uac8c \ubcc0\ud654\ud558\ub294 \uc2dc\uc7a5 \ud658\uacbd\uc5d0 \ub300\uc751\ud558\uae30 \uc704\ud574 \ub370\uc774\ud130 \ubd84\uc11d \ub2a5\ub825\uc744 \ub04c\uc5b4\uc62c\ub9ac\uace0\uc790 \uba38\uc2e0\ub7ec\ub2dd \uae30\ubc95\uc744 \ub3c4\uc785\ud558\ub294 \uac83\uc774 \ud544\uc218\uc801\uc785\ub2c8\ub2e4. \ubcf8 \uac15\uc88c\uc5d0\uc11c\ub294 SEC(\uc99d\uad8c\uac70\ub798\uc704\uc6d0\ud68c)\uc758 \uacf5\uc2dc \ubb38\uc11c\ub97c \ud65c\uc6a9\ud558\uc5ec Word2Vec \uae30\ubc95\uc744 \ud1b5\ud574 \ud14d\uc2a4\ud2b8 \ub370\uc774\ud130\ub97c \ubca1\ud130\ud654\ud558\uace0, \uc774\ub97c \ud1b5\ud574 \uc54c\uace0\ub9ac\uc998 \ud2b8\ub808\uc774\ub529\uc5d0 \uc801\uc6a9\ud558\ub294 \ubc29\ubc95\uc744 \uc790\uc138\ud788 \ub2e4\ub8e8\uaca0\uc2b5\ub2c8\ub2e4.<\/p>\n
\uc8fc\uc2dd \uc2dc\uc7a5\uc5d0\uc11c\uc758 \uc815\ubcf4 \ube44\ub300\uce6d\uc131\uc740 \ud22c\uc790\uc790\uc5d0\uac8c \ud070 \uc704\ud611\uc774 \ub420 \uc218 \uc788\uc2b5\ub2c8\ub2e4. \uacf5\uc2dc \ubb38\uc11c\ub294 \uae30\uc5c5\uc758 \uc7ac\ubb34 \uc0c1\ud0dc, \uacbd\uc601 \uc804\ub7b5 \ubc0f \uc6b4\uc601 \uacb0\uacfc \ub4f1 \ud575\uc2ec \uc815\ubcf4\ub97c \ub2f4\uace0 \uc788\uc73c\uba70, \uc774\ub97c \uae30\ubc18\uc73c\ub85c \ud55c \ubd84\uc11d\uc740 \ud22c\uc790 \uacb0\uc815\uc744 \ub0b4\ub9ac\ub294 \ub370 \uc788\uc5b4 \uc911\uc694\ud55c \uc694\uc18c\uac00 \ub429\ub2c8\ub2e4. \ud558\uc9c0\ub9cc \uc774\ub7ec\ud55c \ubc29\ub300\ud55c \uc591\uc758 \ud14d\uc2a4\ud2b8 \ub370\uc774\ud130\ub97c \uc218\uc791\uc5c5\uc73c\ub85c \ubd84\uc11d\ud558\uae30\ub294 \ubd88\uac00\ub2a5\ud569\ub2c8\ub2e4. \ub530\ub77c\uc11c, \uba38\uc2e0\ub7ec\ub2dd\uacfc \ub525\ub7ec\ub2dd \uae30\ubc95\uc744 \uc801\uc6a9\ud558\uc5ec \ubb38\uc790 \ub370\uc774\ud130\ub97c \uad6c\uc870\ud654\ub41c \ud615\ud0dc\ub85c \ubcc0\ud658\ud558\uace0 \uc774\ub97c \ud2b8\ub808\uc774\ub529 \uc804\ub7b5\uc5d0 \ud65c\uc6a9\ud560 \uc218 \uc788\ub294 \ubc29\ubc95\ub860\uc744 \uc81c\uc2dc\ud560 \uac83\uc785\ub2c8\ub2e4.<\/p>\n
SEC\ub294 \ubbf8\uad6d\uc758 \uc99d\uad8c \uc2dc\uc7a5\uc5d0\uc11c \ud22c\uc790\uc790 \ubcf4\ud638\uc640 \uc2dc\uc7a5\uc758 \uacf5\uc815\uc131\uc744 \ud655\ubcf4\ud558\uae30 \uc704\ud574 \uae30\uc5c5\uc774 \uc815\uae30\uc801\uc73c\ub85c \uc81c\ucd9c\ud574\uc57c \ud558\ub294 \ubcf4\uace0\uc11c\ub97c \uad00\ub9ac\ud569\ub2c8\ub2e4. \uac00\uc7a5 \uc77c\ubc18\uc801\uc778 \ubcf4\uace0\uc11c\ub294 10-K(\uc5f0\uac04 \ubcf4\uace0\uc11c)\uc640 10-Q(\ubd84\uae30 \ubcf4\uace0\uc11c)\uc785\ub2c8\ub2e4. \uc774\ub7ec\ud55c \ubb38\uc11c\ub4e4\uc740 \ub2e4\uc74c\uacfc \uac19\uc740 \uc815\ubcf4\ub4e4\uc744 \ud3ec\ud568\ud558\uace0 \uc788\uc2b5\ub2c8\ub2e4:<\/p>\n
SEC\uc758 \uacf5\uc2dc \ubb38\uc11c\ub294 EDGAR \uc2dc\uc2a4\ud15c\uc744 \ud1b5\ud574 \uc628\ub77c\uc778\uc5d0\uc11c \uc5f4\ub78c\ud560 \uc218 \uc788\uc73c\uba70, Python\uc758 \uc5ec\ub7ec \ub77c\uc774\ube0c\ub7ec\ub9ac\ub97c \uc774\uc6a9\ud558\uc5ec \ub370\uc774\ud130\ub97c \uc218\uc9d1\ud560 \uc218 \uc788\uc2b5\ub2c8\ub2e4. \uc608\ub97c \ub4e4\uc5b4, `requests`\uc640 `BeautifulSoup` \ub77c\uc774\ube0c\ub7ec\ub9ac\ub97c \ud65c\uc6a9\ud558\uc5ec 10-K \ubcf4\uace0\uc11c\ub97c \ub2e4\uc6b4\ub85c\ub4dc\ud558\uace0, \ud544\uc694\ud55c \uc815\ubcf4\ub97c \ucd94\ucd9c\ud560 \uc218 \uc788\uc2b5\ub2c8\ub2e4.<\/p>\n
import requests\nfrom bs4 import BeautifulSoup\n\ndef download_report(cik):\n # SEC EDGAR \uac80\uc0c9 URL\n url = f'https:\/\/www.sec.gov\/cgi-bin\/browse-edgar?cik={cik}&action=getcompany'\n response = requests.get(url)\n soup = BeautifulSoup(response.content, 'html.parser')\n # \ubb38\uc11c \ub9c1\ud06c \ucc3e\uae30\n links = soup.find_all('a', href=True)\n for link in links:\n if '10-K' in link.text:\n report_link = link['href']\n break\n return report_link\n<\/code><\/pre>\n3. Word2Vec\uc758 \uc774\ud574\uc640 \uad6c\ud604<\/h2>\n
Word2Vec\ub294 \ub2e8\uc5b4\ub97c \uace0\ucc28\uc6d0\uc758 \ubca1\ud130 \uacf5\uac04\uc73c\ub85c \ubcc0\ud658\ud558\ub294 \ub300\ud45c\uc801\uc778 \uc790\uc5f0\uc5b4 \ucc98\ub9ac(NLP) \uae30\uc220\uc785\ub2c8\ub2e4. \uc774 \uae30\ubc95\uc744 \uc0ac\uc6a9\ud558\uba74 \ub2e8\uc5b4\uc758 \uc758\ubbf8\uc640 \ubb38\ub9e5\uc744 \uace0\ub824\ud558\uc5ec \uc720\uc0ac\ud55c \uc758\ubbf8\ub97c \uac00\uc9c4 \ub2e8\uc5b4\ub4e4\uc774 \ube44\uc2b7\ud55c \ubca1\ud130\ub85c \ubc30\uce58\ub429\ub2c8\ub2e4. Word2Vec\ub294 \ub450 \uac00\uc9c0 \ubaa8\ub378\uc778 Continuous Bag of Words(CBOW)\uc640 Skip-Gram\uc744 \uae30\ubc18\uc73c\ub85c \uc791\ub3d9\ud569\ub2c8\ub2e4.<\/p>\n
3.1 \ubaa8\ub378\uc758 \uc6d0\ub9ac<\/h3>\n
CBOW \ubaa8\ub378\uc740 \uc8fc\ubcc0 \ub2e8\uc5b4\ub97c \uae30\ubc18\uc73c\ub85c \uc911\uc2ec \ub2e8\uc5b4\ub97c \uc608\uce21\ud558\uace0, Skip-Gram \ubaa8\ub378\uc740 \uc911\uc2ec \ub2e8\uc5b4\ub97c \uae30\ubc18\uc73c\ub85c \uc8fc\ubcc0 \ub2e8\uc5b4\ub4e4\uc744 \uc608\uce21\ud569\ub2c8\ub2e4. \uc608\ub97c \ub4e4\uc5b4, \ubb38\uc7a5 “I love machine learning”\uc5d0\uc11c “love”\ub77c\ub294 \ub2e8\uc5b4\uac00 \uc911\uc2ec\ub2e8\uc5b4\ub77c\uba74 \uc8fc\ubcc0 \ub2e8\uc5b4\ub4e4\uc740 “I”, “machine”, “learning”\uc774 \ub420 \uac83\uc785\ub2c8\ub2e4.<\/p>\n
3.2 Word2Vec \uad6c\ud604<\/h3>\n
Word2Vec \uad6c\ud604\uc740 `gensim` \ub77c\uc774\ube0c\ub7ec\ub9ac\ub97c \ud1b5\ud574 \uc27d\uac8c \uc218\ud589\ud560 \uc218 \uc788\uc2b5\ub2c8\ub2e4. \ud14d\uc2a4\ud2b8 \ub370\uc774\ud130\ub97c \uc804\ucc98\ub9ac\ud55c \ud6c4, \ubaa8\ub378\uc744 \ud559\uc2b5\ud558\ub294 \uacfc\uc815\uc744 \uc0b4\ud3b4\ubcf4\ub3c4\ub85d \ud558\uaca0\uc2b5\ub2c8\ub2e4.<\/p>\n
from gensim.models import Word2Vec\nfrom nltk.tokenize import word_tokenize\nimport nltk\n\n# nltk\uc758 punkt \ud328\ud0a4\uc9c0 \ub2e4\uc6b4\ub85c\ub4dc\nnltk.download('punkt')\n\n# \ud14d\uc2a4\ud2b8 \ub370\uc774\ud130 \uc804\ucc98\ub9ac \ud568\uc218\ndef preprocess_text(text):\n tokens = word_tokenize(text.lower())\n return tokens\n\n# \uc608\uc81c \ud14d\uc2a4\ud2b8\nexample_text = \"The company reported a significant increase in revenue.\"\n\n# \uc804\ucc98\ub9ac \ubc0f \ubaa8\ub378 \ud559\uc2b5\ntokens = preprocess_text(example_text)\nmodel = Word2Vec([tokens], vector_size=100, window=5, min_count=1, sg=0)\n<\/code><\/pre>\n4. SEC \uacf5\uc2dc \ub370\uc774\ud130 \ud65c\uc6a9\ud558\uae30<\/h2>\n
Word2Vec \ubaa8\ub378\uc744 \ud1b5\ud574 \ubca1\ud130\ud654\ud55c SEC\uc758 \uacf5\uc2dc \ud14d\uc2a4\ud2b8 \ub370\uc774\ud130\ub97c \ubc14\ud0d5\uc73c\ub85c, \uc8fc\uc2dd \uc2dc\uc7a5\uc5d0\uc11c\uc758 \uc608\uce21 \ubaa8\ub378\uc744 \uad6c\ucd95\ud560 \uc218 \uc788\uc2b5\ub2c8\ub2e4. \uc608\ub97c \ub4e4\uc5b4, \ud2b9\uc815 \uae30\uc5c5\uc758 \uacf5\uc2dc\ub0b4\uc6a9\uc744 \ubd84\uc11d\ud558\uc5ec \uc8fc\uac00 \ubcc0\ub3d9\uc744 \uc608\uce21\ud558\ub294 \ubc29\ubc95\uc744 \uc801\uc6a9\ud560 \uc218 \uc788\uc2b5\ub2c8\ub2e4.<\/p>\n
4.1 \ud2b8\ub808\uc774\ub529 \uc2e0\ud638 \uc0dd\uc131<\/h3>\n
\ubca1\ud130\ud654\ud55c \ub370\uc774\ud130\ub97c \ubc14\ud0d5\uc73c\ub85c \uba38\uc2e0\ub7ec\ub2dd \uae30\ubc95\uc744 \ud65c\uc6a9\ud558\uc5ec \ud2b8\ub808\uc774\ub529 \uc2e0\ud638\ub97c \uc0dd\uc131\ud569\ub2c8\ub2e4. \uc9c0\uc6d0 \ubca1\ud130 \uba38\uc2e0(SVM), \ub79c\ub364 \ud3ec\ub808\uc2a4\ud2b8, XGBoost \ub4f1 \ub2e4\uc591\ud55c \uba38\uc2e0\ub7ec\ub2dd \uc54c\uace0\ub9ac\uc998\uc744 \uc120\ud0dd\ud560 \uc218 \uc788\uc2b5\ub2c8\ub2e4. \uac01 \uc54c\uace0\ub9ac\uc998\uc758 \uc131\ub2a5\uc744 \ube44\uad50\ud558\ub294 \uac83\uc740 \uc911\uc694\ud55c \uacfc\uc815\uc785\ub2c8\ub2e4.<\/p>\n
4.1.1 \ub370\uc774\ud130\uc14b \ubd84\ub9ac<\/h4>\n
\ub370\uc774\ud130\uc14b\uc744 \ud559\uc2b5 \ub370\uc774\ud130\uc640 \ud14c\uc2a4\ud2b8 \ub370\uc774\ud130\ub85c \ubd84\ub9ac\ud558\ub294 \uac83\uc774 \uc911\uc694\ud569\ub2c8\ub2e4. \uc77c\ubc18\uc801\uc73c\ub85c 70%\uc5d0\uc11c 80%\ub97c \ud559\uc2b5 \ub370\uc774\ud130\ub85c \uc0ac\uc6a9\ud558\uace0, \ub098\uba38\uc9c0\ub97c \ud14c\uc2a4\ud2b8 \ub370\uc774\ud130\ub85c \uc0ac\uc6a9\ud569\ub2c8\ub2e4.<\/p>\n
from sklearn.model_selection import train_test_split\n\n# \uc608\uc81c \ub370\uc774\ud130\uc14b\nX = [...] # \ubca1\ud130\ud654\ub41c \uc785\ub825 \ub370\uc774\ud130\ny = [...] # \ub300\uc751\ud558\ub294 \ub808\uc774\ube14 (\uc608: \uc8fc\uac00 \uc0c1\uc2b9\/\ud558\ub77d)\nX_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)\n<\/code><\/pre>\n4.1.2 \uba38\uc2e0\ub7ec\ub2dd \ubaa8\ub378 \ud6c8\ub828<\/h4>\nfrom sklearn.ensemble import RandomForestClassifier\nfrom sklearn.metrics import accuracy_score\n\n# \ub79c\ub364 \ud3ec\ub808\uc2a4\ud2b8 \ubaa8\ub378 \ud559\uc2b5\nmodel = RandomForestClassifier()\nmodel.fit(X_train, y_train)\n\n# \ud14c\uc2a4\ud2b8 \ub370\uc774\ud130 \uc608\uce21\ny_pred = model.predict(X_test)\n\n# \uc815\ud655\ub3c4 \ud3c9\uac00\naccuracy = accuracy_score(y_test, y_pred)\nprint(f\"Accuracy: {accuracy * 100:.2f}%\")\n<\/code><\/pre>\n5. \uacb0\uacfc \ubd84\uc11d \ubc0f \uc2dc\uac01\ud654<\/h2>\n
\ud6c8\ub828\ub41c \ubaa8\ub378\uc758 \uc608\uce21 \uacb0\uacfc\ub97c \ubd84\uc11d\ud558\uace0 \uc2dc\uac01\ud654\ud558\ub294 \uac83\uc740 \ubaa8\ub378\uc758 \uc131\ub2a5 \ud3c9\uac00\uc5d0 \ud544\uc218\uc801\uc785\ub2c8\ub2e4. \uc774\ub97c \ud1b5\ud574 \ubaa8\ud615\uc758 \uc720\ud6a8\uc131\uc744 \ud3c9\uac00\ud558\uace0, \ud22c\uc790 \uc804\ub7b5\uc744 \uc218\uc815\ud560 \uc218 \uc788\uc2b5\ub2c8\ub2e4.<\/p>\n
5.1 \ud63c\ub3c8 \ud589\ub82c\uacfc \uc815\ud655\ub3c4<\/h3>\nfrom sklearn.metrics import confusion_matrix\nimport matplotlib.pyplot as plt\nimport seaborn as sns\n\n# \ud63c\ub3c8 \ud589\ub82c \uc0dd\uc131\nconf_matrix = confusion_matrix(y_test, y_pred)\n\n# \uc2dc\uac01\ud654\nplt.figure(figsize=(10,7))\nsns.heatmap(conf_matrix, annot=True, fmt='d', cmap='Blues')\nplt.xlabel('Predicted')\nplt.ylabel('Actual')\nplt.title('Confusion Matrix')\nplt.show()\n<\/code><\/pre>\n6. \uacb0\ub860<\/h2>\n
\ubcf8 \uac15\uc88c\uc5d0\uc11c\ub294 SEC \uacf5\uc2dc \ubb38\uc11c\ub97c \ud65c\uc6a9\ud558\uc5ec Word2Vec \uae30\ubc95\uc744 \uc801\uc6a9\ud55c \uba38\uc2e0\ub7ec\ub2dd \ubc0f \ub525\ub7ec\ub2dd \uae30\ubc18\uc758 \uc54c\uace0\ub9ac\uc998 \ud2b8\ub808\uc774\ub529 \ubc29\ubc95\ub860\uc744 \uc18c\uac1c\ud558\uc600\uc2b5\ub2c8\ub2e4. \uc774 \uacfc\uc815\uc5d0\uc11c \ub370\uc774\ud130 \uc218\uc9d1, \ud14d\uc2a4\ud2b8 \uc804\ucc98\ub9ac, \ubca1\ud130\ud654, \ud2b8\ub808\uc774\ub529 \uc2e0\ud638 \uc0dd\uc131 \ubc0f \uc131\ub2a5 \ud3c9\uac00\uc5d0 \ub300\ud55c \ub2e4\uc591\ud55c \uae30\ubc95\uc744 \ub2e4\ub8e8\uc5c8\uc2b5\ub2c8\ub2e4. \uc774\ub7ec\ud55c \uc811\uadfc \ubc29\ubc95\uc744 \ud1b5\ud574 \ud22c\uc790\uc790\ub4e4\uc740 \uc815\ubcf4\ub97c \ub354 \uc798 \ud65c\uc6a9\ud558\uace0, \ub9ac\uc2a4\ud06c\ub97c \uc904\uc77c \uc218 \uc788\ub294 \ubc29\ubc95\uc744 \ubaa8\uc0c9\ud560 \uc218 \uc788\uc2b5\ub2c8\ub2e4.<\/p>\n
\uc55e\uc73c\ub85c\ub3c4 \ub354 \ub9ce\uc740 \ub370\uc774\ud130\uc640 \ub2e4\uc591\ud55c \uc54c\uace0\ub9ac\uc998\uc744 \ud65c\uc6a9\ud558\uc5ec \uc9c0\uc18d\uc801\uc73c\ub85c \ud559\uc2b5\ud558\uace0 \uac1c\uc120\ud574 \ub098\uac00\uc57c \ud560 \uac83\uc785\ub2c8\ub2e4. \uba38\uc2e0\ub7ec\ub2dd\uacfc \ub525\ub7ec\ub2dd \uae30\uc220\uc758 \ubc1c\uc804\uc740 \uc54c\ub85c\ub9ac\uc998 \ud2b8\ub808\uc774\ub529\uc758 \ud328\ub7ec\ub2e4\uc784\uc744 \ubcc0\ud654\uc2dc\ud0a4\uace0 \uc788\uc73c\uba70, \ud22c\uc790\uc758 \uc0c8\ub85c\uc6b4 \uc9c0\ud3c9\uc744 \uc5f4\uc5b4\uc8fc\uace0 \uc788\uc2b5\ub2c8\ub2e4.<\/p>\n