\ucd5c\uadfc \uba87 \ub144\uac04 \uc554\ud638\ud654\ud3d0 \uc2dc\uc7a5\uc740 \ud070 \ubcc0\ub3d9\uc131\uacfc \ub192\uc740 \uc218\uc775\ub960\ub85c \ud2b8\ub808\uc774\ub529\uc758 \uc8fc\uc694 \ub300\uc0c1\uc73c\ub85c \uc790\ub9ac\uc7a1\uc558\uc2b5\ub2c8\ub2e4. \uc774\ub7ec\ud55c \uc2dc\uc7a5\uc5d0\uc11c \uc131\uacf5\uc801\uc73c\ub85c \uac70\ub798\ub97c \ud558\uae30 \uc704\ud574\uc11c\ub294 \ud6a8\uacfc\uc801\uc778 \ud2b8\ub808\uc774\ub529 \uc804\ub7b5\uc744 \uc138\uc6b0\uace0 \uc774\ub97c \ucd5c\uc801\ud654\ud558\ub294 \uacfc\uc815\uc774 \uc911\uc694\ud569\ub2c8\ub2e4. \ubcf8 \uae00\uc5d0\uc11c\ub294 \uba38\uc2e0\ub7ec\ub2dd\uacfc \ub525\ub7ec\ub2dd\uc744 \ud65c\uc6a9\ud574 \uc790\ub3d9\ub9e4\ub9e4 \uc2dc\uc2a4\ud15c\uc5d0\uc11c \ud000\ud2b8 \ud2b8\ub808\uc774\ub529 \uc804\ub7b5\uc758 \ud30c\ub77c\ubbf8\ud130\ub97c \ucd5c\uc801\ud654\ud558\ub294 \ubc29\ubc95\uc5d0 \ub300\ud574 \uc124\uba85\ud558\uaca0\uc2b5\ub2c8\ub2e4.<\/p>\n
\ud000\ud2b8 \ud2b8\ub808\uc774\ub529\uc740 \uc218\ud559\uc801 \ubaa8\ub378\uacfc \uc54c\uace0\ub9ac\uc998\uc744 \ud65c\uc6a9\ud574 \uc790\uc0b0\uc758 \uac00\uaca9 \ubcc0\ub3d9\uc744 \uc608\uce21\ud558\uace0 \uc774\ub97c \uae30\ubc18\uc73c\ub85c \uac70\ub798 \uacb0\uc815\uc744 \ub0b4\ub9ac\ub294 \uae30\ubc95\uc785\ub2c8\ub2e4. \uc774 \uacfc\uc815\uc5d0\uc11c\ub294 \ud1b5\uacc4\uc801 \ubd84\uc11d, \ub370\uc774\ud130 \ub9c8\uc774\ub2dd, \uba38\uc2e0\ub7ec\ub2dd \uae30\uc220\uc774 \ud65c\uc6a9\ub429\ub2c8\ub2e4.<\/p>\n
\uccab \ubc88\uc9f8 \ub2e8\uacc4\ub294 \uc554\ud638\ud654\ud3d0 \uac00\uaca9 \ub370\uc774\ud130\ub97c \uc218\uc9d1\ud558\uace0 \uc804\ucc98\ub9ac\ud558\ub294 \uac83\uc785\ub2c8\ub2e4. \uc77c\ubc18\uc801\uc73c\ub85c API\ub97c \ud1b5\ud574 \uac00\uaca9 \ub370\uc774\ud130\ub97c \uac00\uc838\uc624\uac70\ub098, CSV \ud30c\uc77c \ud615\uc2dd\uc73c\ub85c \ub2e4\uc6b4\ub85c\ub4dc\ud558\uc5ec \uc0ac\uc6a9\ud560 \uc218 \uc788\uc2b5\ub2c8\ub2e4. \uc774\ubc88 \uc608\uc81c\uc5d0\uc11c\ub294 \ube44\ud2b8\ucf54\uc778 \uac00\uaca9 \ub370\uc774\ud130\ub97c \uc0ac\uc6a9\ud569\ub2c8\ub2e4.<\/p>\n
import pandas as pd\nimport requests\n\n# \ube44\ud2b8\ucf54\uc778 \uac00\uaca9 \ub370\uc774\ud130 \uc218\uc9d1\nurl = 'https:\/\/api.coindesk.com\/v1\/bpi\/historical\/close.json'\ndata = requests.get(url).json()\nbitcoin_prices = pd.DataFrame(data['bpi']).reset_index()\nbitcoin_prices.columns = ['Date', 'Close']\nbitcoin_prices['Date'] = pd.to_datetime(bitcoin_prices['Date'])\n<\/code><\/pre>\n2.2 \ub370\uc774\ud130 \uc804\ucc98\ub9ac<\/h3>\n
\ub370\uc774\ud130\ub97c \uba38\uc2e0\ub7ec\ub2dd \ubaa8\ub378\uc5d0 \ub9de\uac8c \uc804\ucc98\ub9ac\ud569\ub2c8\ub2e4. \uacb0\uce21\uce58\ub97c \ucc98\ub9ac\ud558\uace0, \ud544\uc694\ud560 \uacbd\uc6b0 \uae30\uc220\uc801 \uc9c0\ud45c\ub97c \ucd94\uac00\ub85c \uacc4\uc0b0\ud560 \uc218 \uc788\uc2b5\ub2c8\ub2e4.<\/p>\n
# \uacb0\uce21\uce58 \ucc98\ub9ac\nbitcoin_prices.dropna(inplace=True)\n\n# \uae30\uc220\uc801 \uc9c0\ud45c \ucd94\uac00: \uc774\ub3d9 \ud3c9\uade0 \uacc4\uc0b0\nbitcoin_prices['MA_20'] = bitcoin_prices['Close'].rolling(window=20).mean()\nbitcoin_prices['MA_50'] = bitcoin_prices['Close'].rolling(window=50).mean()\n<\/code><\/pre>\n3. \uba38\uc2e0\ub7ec\ub2dd \ubaa8\ub378 \uad6c\ucd95<\/h2>\n
\uc900\ube44\ud55c \ub370\uc774\ud130\ub97c \ubc14\ud0d5\uc73c\ub85c \uba38\uc2e0\ub7ec\ub2dd \ubaa8\ub378\uc744 \ud6c8\ub828\ud569\ub2c8\ub2e4. \ud68c\uadc0 \ub610\ub294 \ubd84\ub958 \ubb38\uc81c\ub97c \uc815\uc758\ud558\uace0 \uc0ac\uc6a9\ud558\ub294 \uc54c\uace0\ub9ac\uc998\uc744 \uc120\ud0dd\ud569\ub2c8\ub2e4. \uc77c\ubc18\uc801\uc73c\ub85c \uc0ac\uc6a9\ub418\ub294 \uc54c\uace0\ub9ac\uc998\uc73c\ub85c\ub294 \ub79c\ub364 \ud3ec\ub808\uc2a4\ud2b8, SVM, LSTM \ub4f1\uc774 \uc788\uc2b5\ub2c8\ub2e4.<\/p>\n
3.1 \ub370\uc774\ud130\uc14b \ubd84\ub9ac<\/h3>\nfrom sklearn.model_selection import train_test_split\n\n# Feature\uc640 Target \uc815\uc758\nX = bitcoin_prices[['MA_20', 'MA_50']].dropna()\ny = bitcoin_prices['Close'].shift(-1).dropna()\n\n# \ub370\uc774\ud130\uc14b \ubd84\ub9ac\nX_train, X_test, y_train, y_test = train_test_split(X[:-1], y, test_size=0.2, random_state=42)\n<\/code><\/pre>\n3.2 \ubaa8\ub378 \ud6c8\ub828<\/h3>\nfrom sklearn.ensemble import RandomForestRegressor\n\n# \uba38\uc2e0\ub7ec\ub2dd \ubaa8\ub378 \uc815\uc758 \ubc0f \ud6c8\ub828\nmodel = RandomForestRegressor(n_estimators=100)\nmodel.fit(X_train, y_train)\n<\/code><\/pre>\n4. \uc804\ub7b5 \ucd5c\uc801\ud654<\/h2>\n
\uc774\uc81c \ubaa8\ub378\uc774 \ud6c8\ub828\ub418\uc5c8\uc73c\ubbc0\ub85c, \ub2e4\uc74c \ub2e8\uacc4\ub294 \ubaa8\ub378\uc758 \ud558\uc774\ud37c\ud30c\ub77c\ubbf8\ud130\ub97c \ucd5c\uc801\ud654\ud558\ub294 \uac83\uc785\ub2c8\ub2e4. \uc774\ub97c \uc704\ud574 Grid Search \ub610\ub294 Random Search\uc640 \uac19\uc740 \uae30\ubc95\uc744 \uc0ac\uc6a9\ud560 \uc218 \uc788\uc2b5\ub2c8\ub2e4.<\/p>\n
4.1 \ud558\uc774\ud37c\ud30c\ub77c\ubbf8\ud130 \ucd5c\uc801\ud654<\/h3>\nfrom sklearn.model_selection import GridSearchCV\n\n# \ud30c\ub77c\ubbf8\ud130 \ubc94\uc704 \uc124\uc815\nparam_grid = {\n 'n_estimators': [50, 100, 200],\n 'max_depth': [None, 10, 20, 30],\n 'min_samples_split': [2, 5, 10]\n}\n\n# Grid Search \uc2e4\ud589\ngrid_search = GridSearchCV(estimator=model, param_grid=param_grid, cv=3, n_jobs=-1)\ngrid_search.fit(X_train, y_train)\n\n# \ucd5c\uc801 \ud558\uc774\ud37c\ud30c\ub77c\ubbf8\ud130 \ucd9c\ub825\nbest_params = grid_search.best_params_\nprint(f\"Best parameters: {best_params}\")\n<\/code><\/pre>\n5. \uc131\ub2a5 \ud3c9\uac00<\/h2>\n
\ucd5c\uc801\ud654\ub41c \ubaa8\ub378\uc744 \ud3c9\uac00\ud569\ub2c8\ub2e4. RMSE \ubc0f R\u00b2 \uc9c0\ud45c\ub97c \uc0ac\uc6a9\ud558\uc5ec \ubaa8\ub378\uc758 \uc131\ub2a5\uc744 \uce21\uc815\ud560 \uc218 \uc788\uc2b5\ub2c8\ub2e4.<\/p>\n
from sklearn.metrics import mean_squared_error, r2_score\nimport numpy as np\n\n# \uc608\uce21 \uc218\ud589\ny_pred = grid_search.predict(X_test)\n\n# \uc131\ub2a5 \ud3c9\uac00\nrmse = np.sqrt(mean_squared_error(y_test, y_pred))\nr2 = r2_score(y_test, y_pred)\n\nprint(f\"RMSE: {rmse}, R\u00b2: {r2}\")\n<\/code><\/pre>\n6. \uc790\ub3d9\ub9e4\ub9e4 \uc2dc\uc2a4\ud15c \uad6c\ucd95<\/h2>\n
\ubaa8\ub378\uc744 \ubc14\ud0d5\uc73c\ub85c \uc2e4\uc81c \ub9e4\ub9e4\ub97c \uc218\ud589\ud560 \uc218 \uc788\ub294 \uc2dc\uc2a4\ud15c\uc744 \uad6c\ucd95\ud569\ub2c8\ub2e4. \uc774 \ubd80\ubd84\uc740 \uc8fc\uae30\uc801\uc73c\ub85c \uac00\uaca9 \ub370\uc774\ud130\ub97c \uac00\uc838\uc624\uace0 \uc608\uce21\uc744 \uc218\ud589\ud558\uc5ec \ub9e4\ub9e4\uc2e0\ud638\ub97c \uacb0\uc815\ud558\ub294 \ub85c\uc9c1\uc774 \ud3ec\ud568\ub429\ub2c8\ub2e4.<\/p>\n
6.1 \ub9e4\ub9e4 \ub85c\uc9c1 \uad6c\ud604<\/h3>\ndef trade_signal(model, new_data):\n prediction = model.predict(new_data)\n if prediction > new_data['Close'].values[-1]:\n return \"BUY\"\n else:\n return \"SELL\"\n\n# \uc0c8\ub85c\uc6b4 \ub370\uc774\ud130\uac00 \ub4e4\uc5b4\uc62c \ub54c\ub9c8\ub2e4 trade_signal \ud638\ucd9c\nnew_data = bitcoin_prices.iloc[-1][['MA_20', 'MA_50']].values.reshape(1, -1)\nprint(f\"Trade Signal: {trade_signal(grid_search.best_estimator_, new_data)}\")\n<\/code><\/pre>\n7. \uacb0\ub860<\/h2>\n
\ub525\ub7ec\ub2dd\uacfc \uba38\uc2e0\ub7ec\ub2dd\uc744 \ud65c\uc6a9\ud55c \uc790\ub3d9\ub9e4\ub9e4 \uc2dc\uc2a4\ud15c\uc740 \ub192\uc740 \uc218\uc775\uc744 \uac00\uc838\ub2e4 \uc904 \uc218 \uc788\ub294 \uc7a0\uc7ac\ub825\uc744 \uac00\uc9c0\uace0 \uc788\uc2b5\ub2c8\ub2e4. \ubcf8 \uae00\uc5d0\uc11c\ub294 \ub370\uc774\ud130 \uc218\uc9d1\uc5d0\uc11c\ubd80\ud130 \ubaa8\ub378 \ud6c8\ub828, \ucd5c\uc801\ud654 \ubc0f \uc790\ub3d9\ub9e4\ub9e4 \uc2dc\uc2a4\ud15c \uad6c\ucd95\uc5d0 \uc774\ub974\ub294 \uc804\uccb4 \uacfc\uc815\uc744 \ub2e4\ub8e8\uc5c8\uc2b5\ub2c8\ub2e4. \uc774\ub7ec\ud55c \uc2dc\uc2a4\ud15c\uc744 \ud1b5\ud574 \ud000\ud2b8 \ud2b8\ub808\uc774\ub529 \uc804\ub7b5\uc744 \ud55c\uce35 \uac15\ud654\ud560 \uc218 \uc788\uc744 \uac83\uc785\ub2c8\ub2e4.<\/p>\n
8. References<\/h2>\n