{"id":31379,"date":"2024-10-28T12:01:38","date_gmt":"2024-10-28T12:01:38","guid":{"rendered":"http:\/\/atmokpo.com\/w\/?p=31379"},"modified":"2024-11-26T06:45:21","modified_gmt":"2024-11-26T06:45:21","slug":"%eb%94%a5%eb%9f%ac%eb%8b%9d%ea%b3%bc-%eb%a8%b8%ec%8b%a0%eb%9f%ac%eb%8b%9d%ec%9d%84-%ed%99%9c%ec%9a%a9%ed%95%9c-%ec%9e%90%eb%8f%99%eb%a7%a4%eb%a7%a4-%eb%93%9c%eb%a1%ad%ec%95%84%ec%9b%83%ea%b3%bc","status":"publish","type":"post","link":"https:\/\/atmokpo.com\/w\/31379\/","title":{"rendered":"\ub525\ub7ec\ub2dd\uacfc \uba38\uc2e0\ub7ec\ub2dd\uc744 \ud65c\uc6a9\ud55c \uc790\ub3d9\ub9e4\ub9e4, \ub4dc\ub86d\uc544\uc6c3\uacfc \ubc30\uce58 \uc815\uaddc\ud654\ub97c \ud65c\uc6a9\ud55c \ubaa8\ub378 \uc548\uc815\ud654 \ub525\ub7ec\ub2dd \ubaa8\ub378\uc758 \uc548\uc815\uc801\uc778 \ud559\uc2b5\uc744 \uc704\ud55c \uae30\uc220 \uc801\uc6a9."},"content":{"rendered":"

<\/p>\n

\ube44\ud2b8\ucf54\uc778\uacfc \uac19\uc740 \uc554\ud638\ud654\ud3d0 \uc2dc\uc7a5\uc740 \ubcc0\ub3d9\uc131\uc774 \ud06c\uace0, \ubd88\ud655\uc2e4\uc131\uc774 \uc874\uc7ac\ud569\ub2c8\ub2e4. \uc774\ub7ec\ud55c \uc2dc\uc7a5\uc5d0\uc11c \uc790\ub3d9\ub9e4\ub9e4 \uc2dc\uc2a4\ud15c\uc744 \uad6c\ucd95\ud558\uae30 \uc704\ud574\uc11c\ub294 \ub525\ub7ec\ub2dd\uacfc \uba38\uc2e0\ub7ec\ub2dd \uae30\ubc95\uc744 \uc801\uc6a9\ud558\uc5ec \ud6a8\uacfc\uc801\uc778 \uc608\uce21 \ubaa8\ub378\uc744 \ub9cc\ub4e4\uc5b4\uc57c \ud569\ub2c8\ub2e4. \ubcf8 \uac15\uc88c\uc5d0\uc11c\ub294 \ub4dc\ub86d\uc544\uc6c3(Dropout)<\/span>\uacfc \ubc30\uce58 \uc815\uaddc\ud654(Batch Normalization)<\/span> \uac19\uc740 \uae30\uc220\uc744 \ud65c\uc6a9\ud558\uc5ec \ubaa8\ub378\uc758 \uc548\uc815\uc131\uc744 \ub192\uc774\uace0 \uc131\ub2a5\uc744 \ud5a5\uc0c1\uc2dc\ud0a4\ub294 \ubc29\ubc95\uc5d0 \ub300\ud574 \uc54c\uc544\ubcf4\uaca0\uc2b5\ub2c8\ub2e4.<\/p>\n

1. \uc790\ub3d9\ub9e4\ub9e4 \uc2dc\uc2a4\ud15c\uc758 \uac1c\uc694<\/h2>\n

\uc790\ub3d9\ub9e4\ub9e4 \uc2dc\uc2a4\ud15c\uc740 \uc54c\uace0\ub9ac\uc998\uc744 \uae30\ubc18\uc73c\ub85c \ub9e4\ub9e4 \uacb0\uc815\uc744 \ub0b4\ub9ac\ub294 \uc2dc\uc2a4\ud15c\uc785\ub2c8\ub2e4. \uc774\ub7f0 \uc2dc\uc2a4\ud15c\uc740 \uac00\uaca9, \uac70\ub798\ub7c9, \uae30\uc220\uc801 \uc9c0\ud45c \ub4f1 \ub2e4\uc591\ud55c \ub370\uc774\ud130\ub97c \ubd84\uc11d\ud558\uc5ec \ub9e4\uc218 \ud639\uc740 \ub9e4\ub3c4 \uc2e0\ud638\ub97c \ubc1c\uc0dd\uc2dc\ud0b5\ub2c8\ub2e4. \uba38\uc2e0\ub7ec\ub2dd \ubc0f \ub525\ub7ec\ub2dd \uae30\ubc95\uc744 \uc0ac\uc6a9\ud558\uc5ec \ub354 \ud5a5\uc0c1\ub41c \uc608\uce21\ub825\uc744 \ubc1c\ud718\ud560 \uc218 \uc788\uc2b5\ub2c8\ub2e4.<\/p>\n

1.1. \ub370\uc774\ud130 \uc218\uc9d1<\/h3>\n

\uc790\ub3d9\ub9e4\ub9e4\ub97c \uc704\ud55c \uccab \ubc88\uc9f8 \ub2e8\uacc4\ub294 \ub370\uc774\ud130 \uc218\uc9d1\uc785\ub2c8\ub2e4. \ube44\ud2b8\ucf54\uc778 \uac00\uaca9 \ub370\uc774\ud130\ub294 \ub2e4\uc591\ud55c API\ub97c \ud1b5\ud574 \uc218\uc9d1\ud560 \uc218 \uc788\uc2b5\ub2c8\ub2e4. \uc608\ub97c \ub4e4\uc5b4, CoinGecko<\/code>\ub098 Binance API<\/code>\ub97c \uc0ac\uc6a9\ud574 BTC-USD \uac00\uaca9 \ub370\uc774\ud130\ub97c \uac00\uc838\uc62c \uc218 \uc788\uc2b5\ub2c8\ub2e4. \uc218\uc9d1\ud558\ub294 \ub370\uc774\ud130\ub294 \uc2dc\uac04, \uac00\uaca9, \uac70\ub798\ub7c9 \ub4f1 \ub2e4\uc591\ud55c \uc694\uc18c\ub97c \ud3ec\ud568\ud574\uc57c \ud569\ub2c8\ub2e4.<\/p>\n

1.2. \ub370\uc774\ud130 \uc804\ucc98\ub9ac<\/h3>\n

\uc218\uc9d1\ud55c \ub370\uc774\ud130\ub294 \ubaa8\ub378\uc5d0 \uc785\ub825\ud558\uae30 \uc804\uc5d0 \uc804\ucc98\ub9ac\ub97c \uac70\uccd0\uc57c \ud569\ub2c8\ub2e4. \uacb0\uce21\uce58 \ucc98\ub9ac, \uc815\uaddc\ud654(Normalization), \uadf8\ub9ac\uace0 \ud2b9\uc9d5 \uc120\ud0dd(Feature Selection) \ub4f1\uc758 \uacfc\uc815\uc774 \ud544\uc694\ud569\ub2c8\ub2e4. \uc608\ub97c \ub4e4\uc5b4, \uc885\uac00\ub97c \uc774\uc6a9\ud558\uc5ec \uac04\ub2e8\ud55c \uc815\uaddc\ud654\ub97c \uc9c4\ud589\ud560 \uc218 \uc788\uc2b5\ub2c8\ub2e4.<\/p>\n

2. \ub525\ub7ec\ub2dd \ubaa8\ub378\uc758 \uad6c\ucd95<\/h2>\n

\ub525\ub7ec\ub2dd \ubaa8\ub378\uc744 \uad6c\ucd95\ud558\uae30 \uc704\ud574\uc11c\ub294 TensorFlow, Keras \ub610\ub294 PyTorch \uac19\uc740 \ub77c\uc774\ube0c\ub7ec\ub9ac\ub97c \uc0ac\uc6a9\ud560 \uc218 \uc788\uc2b5\ub2c8\ub2e4. \ubcf8 \uc608\uc81c\uc5d0\uc11c\ub294 Keras\ub97c \uc0ac\uc6a9\ud558\uc5ec \uac04\ub2e8\ud55c \ubaa8\ub378\uc744 \ub9cc\ub4e4\uc5b4 \ubcf4\uaca0\uc2b5\ub2c8\ub2e4.<\/p>\n

\nimport numpy as np\nimport pandas as pd\nfrom sklearn.preprocessing import MinMaxScaler\nfrom keras.models import Sequential\nfrom keras.layers import Dense, Dropout, BatchNormalization\nfrom sklearn.model_selection import train_test_split\n\n# \ub370\uc774\ud130 \ub85c\ub529 \ubc0f \uc804\ucc98\ub9ac\ndata = pd.read_csv('bitcoin_price.csv')  # \ub370\uc774\ud130 \ud30c\uc77c\ndata['Close'] = data['Close'].shift(-1)  # \ub2e4\uc74c \ub0a0 \uc885\uac00 \uc608\uce21\ndata.dropna(inplace=True)\n\n# \uc815\uaddc\ud654\nscaler = MinMaxScaler()\nscaled_data = scaler.fit_transform(data[['Close', 'Volume']])\nX = scaled_data[:-1]\ny = scaled_data[1:, 0]  # \ub2e4\uc74c \ub0a0 \uc885\uac00\n\n# \ub370\uc774\ud130 \ubd84\ud560\nX_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)\n    <\/code><\/pre>\n
3. \ubaa8\ub378 \uc548\uc815\ud654\ub97c \uc704\ud55c \ub4dc\ub86d\uc544\uc6c3\uacfc \ubc30\uce58 \uc815\uaddc\ud654<\/h2>\n
\ub525\ub7ec\ub2dd \ubaa8\ub378\uc758 \ud559\uc2b5 \uacfc\uc815\uc5d0\uc11c \uc624\ubc84\ud53c\ud305(Overfitting) \ubb38\uc81c\ub97c \ubc29\uc9c0\ud558\uae30 \uc704\ud574 \uc5ec\ub7ec\uac00\uc9c0 \uae30\ubc95\uc774 \uc801\uc6a9\ub429\ub2c8\ub2e4. \uadf8\uc911\uc5d0\uc11c \ub4dc\ub86d\uc544\uc6c3\uacfc \ubc30\uce58 \uc815\uaddc\ud654\ub294 \uac00\uc7a5 \ub9ce\uc774 \uc0ac\uc6a9\ub418\ub294 \uae30\ubc95\uc785\ub2c8\ub2e4.<\/p>\n
3.1. \ub4dc\ub86d\uc544\uc6c3 (Dropout)<\/h3>\n
\ub4dc\ub86d\uc544\uc6c3\uc740 \ud559\uc2b5 \uacfc\uc815\uc5d0\uc11c \ud2b9\uc815 \ub274\ub7f0\uc744 \uc784\uc758\ub85c \uc0dd\ub7b5\ud558\uc5ec \ub124\ud2b8\uc6cc\ud06c\uc758 \uc77c\ubc18\ud654\ub97c \uc99d\uac00\uc2dc\ud0a4\ub294 \uae30\ubc95\uc785\ub2c8\ub2e4. \uc774 \ubc29\uc2dd \ub355\ubd84\uc5d0 \ubaa8\ub378\uc740 \ubd88\ud544\uc694\ud55c \ud328\ud134\uc744 \uc778\uc2dd\ud560 \uac00\ub2a5\uc131\uc774 \uc904\uc5b4\ub4e4\uace0, \ub354 \uc77c\ubc18\uc801\uc778 \ud2b9\uc131\uc744 \ud559\uc2b5\ud560 \uc218 \uc788\uc2b5\ub2c8\ub2e4.<\/p>\n
3.2. \ubc30\uce58 \uc815\uaddc\ud654 (Batch Normalization)<\/h3>\n
\ubc30\uce58 \uc815\uaddc\ud654\ub294 \uac01 \ubbf8\ub2c8\ubc30\uce58\uc758 \ud3c9\uade0\uacfc \ubd84\uc0b0\uc744 \uc0ac\uc6a9\ud574 \ub370\uc774\ud130\ub97c \uc815\uaddc\ud654\ud558\ub294 \ubc29\ubc95\uc785\ub2c8\ub2e4. \uc774 \uae30\ubc95\uc740 \ud559\uc2b5 \uc18d\ub3c4\ub97c \uc99d\uac00\uc2dc\ud0a4\uace0, \uc624\ubc84\ud53c\ud305\uc744 \uc904\uc774\ub294 \ub370 \ub3c4\uc6c0\uc744 \uc90d\ub2c8\ub2e4.<\/p>\n
4. \ubaa8\ub378 \uad6c\ucd95 \ubc0f \ud559\uc2b5<\/h2>\n
\n# \ubaa8\ub378 \uad6c\ucd95\nmodel = Sequential()\nmodel.add(Dense(64, activation='relu', input_shape=(X.shape[1],)))\nmodel.add(Dropout(0.5))  # \ub4dc\ub86d\uc544\uc6c3 \uc801\uc6a9\nmodel.add(BatchNormalization())\nmodel.add(Dense(32, activation='relu'))\nmodel.add(Dropout(0.5))  # \ub4dc\ub86d\uc544\uc6c3 \uc801\uc6a9\nmodel.add(BatchNormalization())\nmodel.add(Dense(1, activation='linear'))\n\n# \ubaa8\ub378 \ucef4\ud30c\uc77c\nmodel.compile(optimizer='adam', loss='mean_squared_error')\n\n# \ubaa8\ub378 \ud559\uc2b5\nmodel.fit(X_train, y_train, epochs=50, batch_size=32, validation_split=0.2)\n    <\/code><\/pre>\n
5. \ubaa8\ub378 \ud3c9\uac00 \ubc0f \uc608\uce21<\/h2>\n
\ud559\uc2b5\ub41c \ubaa8\ub378\uc744 \ud3c9\uac00\ud558\uace0, \ud5a5\ud6c4 \uac00\uaca9 \uc608\uce21\uc744 \uc9c4\ud589\ud569\ub2c8\ub2e4. \ubaa8\ub378\uc758 \uc131\ub2a5\uc744 \ud3c9\uac00\ud558\uae30 \uc704\ud574 MSE(Mean Squared Error) \ub4f1\uc758 \uc9c0\ud45c\ub97c \uc0ac\uc6a9\ud560 \uc218 \uc788\uc2b5\ub2c8\ub2e4.<\/p>\n
\n# \ubaa8\ub378 \ud3c9\uac00\nloss = model.evaluate(X_test, y_test)\nprint(f'\ud14c\uc2a4\ud2b8 \uc190\uc2e4: {loss}')\n\n# \uc608\uce21\npredictions = model.predict(X_test)\npredictions = scaler.inverse_transform(predictions)  # \uc5ed\uc815\uaddc\ud654\n    <\/code><\/pre>\n
6. \uacb0\ub860<\/h2>\n
\ub525\ub7ec\ub2dd\uacfc \uba38\uc2e0\ub7ec\ub2dd \uae30\ubc95\uc744 \ud65c\uc6a9\ud55c \ube44\ud2b8\ucf54\uc778 \uc790\ub3d9\ub9e4\ub9e4 \uc2dc\uc2a4\ud15c\uc740 \ubcc0\ud654\ud558\ub294 \uc2dc\uc7a5\uc5d0\uc11c \ubcf4\ub2e4 \ud6a8\uacfc\uc801\uc778 \uc758\uc0ac\uacb0\uc815\uc744 \uac00\ub2a5\ud558\uac8c \ud569\ub2c8\ub2e4. \ub4dc\ub86d\uc544\uc6c3\uacfc \ubc30\uce58 \uc815\uaddc\ud654 \ub4f1\uc758 \uae30\ubc95\uc744 \ud1b5\ud574 \ubaa8\ub378\uc758 \uc548\uc815\uc131\uc744 \ub192\uc774\uace0, \uc624\ubc84\ud53c\ud305\uc744 \ubc29\uc9c0\ud558\uc5ec \uc608\uce21 \uc131\ub2a5\uc744 \ud5a5\uc0c1\uc2dc\ud0ac \uc218 \uc788\uc2b5\ub2c8\ub2e4. \ub370\uc774\ud130 \uc218\uc9d1, \uc804\ucc98\ub9ac, \ubaa8\ub378 \uad6c\ucd95, \ud559\uc2b5, \ud3c9\uac00 \ubc0f \uc608\uce21\uc758 \ubaa8\ub4e0 \uacfc\uc815\uc740 \ucca0\uc800\ud558\uac8c \uc774\ub8e8\uc5b4\uc838\uc57c \ud558\uba70, \uc9c0\uc18d\uc801\uc778 \ubaa8\ub378 \uac1c\uc120\uc744 \ud1b5\ud574 \ucd5c\uc801\uc758 \uc131\uacfc\ub97c \uac70\ub458 \uc218 \uc788\uc2b5\ub2c8\ub2e4.<\/p>\n
\ucc38\uace0 \uc790\ub8cc<\/h2>\n