{"id":31355,"date":"2024-10-28T12:01:28","date_gmt":"2024-10-28T12:01:28","guid":{"rendered":"http:\/\/atmokpo.com\/w\/?p=31355"},"modified":"2024-11-26T06:45:25","modified_gmt":"2024-11-26T06:45:25","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-hyperparameter-tuning-%eb%94%a5%eb%9f%ac%eb%8b%9d","status":"publish","type":"post","link":"https:\/\/atmokpo.com\/w\/31355\/","title":{"rendered":"\ub525\ub7ec\ub2dd\uacfc \uba38\uc2e0\ub7ec\ub2dd\uc744 \ud65c\uc6a9\ud55c \uc790\ub3d9\ub9e4\ub9e4, Hyperparameter Tuning \ub525\ub7ec\ub2dd \ubaa8\ub378\uc758 \uc131\ub2a5\uc744 \ub192\uc774\uae30 \uc704\ud55c \ud558\uc774\ud37c\ud30c\ub77c\ubbf8\ud130 \ud29c\ub2dd \ubc29\ubc95."},"content":{"rendered":"

<\/p>\n

\uc774\ubc88 \uac15\uc88c\uc5d0\uc11c\ub294 \ub525\ub7ec\ub2dd\uacfc \uba38\uc2e0\ub7ec\ub2dd\uc744 \uc774\uc6a9\ud55c \ube44\ud2b8\ucf54\uc778 \uc790\ub3d9\ub9e4\ub9e4 \uc2dc\uc2a4\ud15c\uc744 \uad6c\ucd95\ud558\ub294 \uacfc\uc815\uc744 \uc0b4\ud3b4\ubcf4\uaca0\uc2b5\ub2c8\ub2e4. \ud2b9\ud788, \uc131\ub2a5\uc744 \uadf9\ub300\ud654\ud558\uae30 \uc704\ud55c \ud558\uc774\ud37c\ud30c\ub77c\ubbf8\ud130 \ud29c\ub2dd\uc758 \uc911\uc694\uc131\uacfc \uadf8 \ubc29\ubc95\uc5d0 \ub300\ud574 \uc0c1\uc138\ud788 \uc124\uba85\ud558\uaca0\uc2b5\ub2c8\ub2e4. \uc6b0\ub9ac\uac00 \uc0ac\uc6a9\ud560 \ub370\uc774\ud130\uc640 \uba38\uc2e0\ub7ec\ub2dd \ubaa8\ub378\uc5d0 \ub300\ud55c \uc18c\uac1c, \uadf8\ub9ac\uace0 \ud558\uc774\ud37c\ud30c\ub77c\ubbf8\ud130 \ud29c\ub2dd \uae30\ubc95\uc744 \uc2e4\uc81c \ucf54\ub4dc \uc608\uc81c\uc640 \ud568\uaed8 \uc81c\uacf5\ud569\ub2c8\ub2e4.<\/p>\n

1. \ube44\ud2b8\ucf54\uc778 \uc790\ub3d9\ub9e4\ub9e4 \uc2dc\uc2a4\ud15c\uc758 \uac1c\uc694<\/h2>\n

\uc790\ub3d9\ub9e4\ub9e4 \uc2dc\uc2a4\ud15c\uc740 \uc8fc\uc2dd, \uc554\ud638\ud654\ud3d0 \ub4f1\uc758 \uc790\uc0b0\uc744 \uac70\ub798\ud558\uae30 \uc704\ud574 \uc0ac\uc6a9\ub418\ub294 \uc54c\uace0\ub9ac\uc998\uc785\ub2c8\ub2e4. \uc774\ub7ec\ud55c \uc2dc\uc2a4\ud15c\uc740 \ub370\uc774\ud130 \ubd84\uc11d, \ud328\ud134 \uc778\uc2dd, \uadf8\ub9ac\uace0 \uc608\uce21 \ubaa8\ub378\ub9c1\uc744 \ud1b5\ud574 \uacb0\uc815\uc744 \ub0b4\ub9bd\ub2c8\ub2e4. \ube44\ud2b8\ucf54\uc778\uc740 \ud2b9\ud788 \ubcc0\ub3d9\uc131\uc774 \ud06c\uae30 \ub54c\ubb38\uc5d0, \uba38\uc2e0\ub7ec\ub2dd \ubc0f \ub525\ub7ec\ub2dd \ubaa8\ub378\uc744 \ud65c\uc6a9\ud558\uc5ec \ud6a8\uacfc\uc801\uc73c\ub85c \ub9e4\ub9e4\ub97c \uc790\ub3d9\ud654\ud560 \uc218 \uc788\uc2b5\ub2c8\ub2e4.<\/p>\n

2. \ud558\uc774\ud37c\ud30c\ub77c\ubbf8\ud130 \ud29c\ub2dd\uc758 \uc911\uc694\uc131<\/h2>\n

\ud558\uc774\ud37c\ud30c\ub77c\ubbf8\ud130\ub294 \uba38\uc2e0\ub7ec\ub2dd \ubaa8\ub378\uc758 \ud559\uc2b5 \uacfc\uc815\uc5d0\uc11c \uc124\uc815\ud574\uc57c \ud558\ub294 \ub9e4\uac1c\ubcc0\uc218\uc785\ub2c8\ub2e4. \uc774\ub4e4\uc740 \ud559\uc2b5\ub960, \ubc30\uce58 \ud06c\uae30, \uc815\uaddc\ud654 \uacc4\uc218 \ub4f1 \ub2e4\uc591\ud558\uba70, \uc774\ub4e4 \uac12\uc5d0 \ub530\ub77c \ubaa8\ub378\uc758 \uc131\ub2a5\uc774 \ud06c\uac8c \ub2ec\ub77c\uc9c8 \uc218 \uc788\uc2b5\ub2c8\ub2e4. \uc801\uc808\ud55c \ud558\uc774\ud37c\ud30c\ub77c\ubbf8\ud130\ub97c \ucc3e\ub294 \uacfc\uc815\uc740 \ubaa8\ub378 \uac1c\uc120\uc758 \uac00\uc7a5 \ud575\uc2ec\uc801\uc778 \ubd80\ubd84 \uc911 \ud558\ub098\uc785\ub2c8\ub2e4.<\/p>\n

3. \ud558\uc774\ud37c\ud30c\ub77c\ubbf8\ud130 \ud29c\ub2dd \uae30\ubc95<\/h2>\n

\ud558\uc774\ud37c\ud30c\ub77c\ubbf8\ud130 \ud29c\ub2dd\uc5d0\ub294 \uc5ec\ub7ec \uac00\uc9c0 \ubc29\ubc95\uc774 \uc788\uc2b5\ub2c8\ub2e4. \uc5ec\uae30\uc11c\ub294 \ub300\ud45c\uc801\uc778 \ub450 \uac00\uc9c0 \ubc29\ubc95\uc778 Grid Search\uc640 Random Search\ub97c \uc18c\uac1c\ud558\uaca0\uc2b5\ub2c8\ub2e4.<\/p>\n

3.1 Grid Search<\/h3>\n

Grid Search\ub294 \ubbf8\ub9ac \uc815\uc758\ub41c \ud558\uc774\ud37c\ud30c\ub77c\ubbf8\ud130 \uac12\uc758 \uc870\ud569\uc744 \ubaa8\ub450 \ud0d0\uc0c9\ud558\uc5ec \ucd5c\uc801\uc758 \uc870\ud569\uc744 \ucc3e\ub294 \ubc29\ubc95\uc785\ub2c8\ub2e4. \uc774 \ubc29\ubc95\uc740 \uac04\ub2e8\ud558\uc9c0\ub9cc \uacc4\uc0b0 \ube44\uc6a9\uc774 \ub9ce\uc774 \ub4e4 \uc218 \uc788\uc2b5\ub2c8\ub2e4.<\/p>\n

from sklearn.model_selection import GridSearchCV\nfrom sklearn.ensemble import RandomForestClassifier\n\n# Hyperparameter grid\nparam_grid = {\n    'n_estimators': [10, 50, 100],\n    'max_features': ['auto', 'sqrt', 'log2'],\n    'max_depth': [None, 10, 20, 30],\n}\n\ngrid_search = GridSearchCV(estimator=RandomForestClassifier(), param_grid=param_grid, cv=3)\ngrid_search.fit(X_train, y_train)\nbest_params = grid_search.best_params_\n<\/code><\/pre>\n
3.2 Random Search<\/h3>\n
Random Search\ub294 \ud558\uc774\ud37c\ud30c\ub77c\ubbf8\ud130 \uacf5\uac04\uc5d0\uc11c \uc784\uc758\uc758 \uc870\ud569\uc744 \uc120\ud0dd\ud558\uc5ec \uc131\ub2a5\uc744 \ud3c9\uac00\ud558\ub294 \ubc29\ubc95\uc785\ub2c8\ub2e4. Grid Search\ubcf4\ub2e4 \ube60\ub974\uac8c \ucd5c\uc801\uc758 \uc870\ud569\uc744 \ucc3e\uc744 \uc218 \uc788\uc9c0\ub9cc, \uc801\uc808\ud55c \uc870\ud569\uc758 \uc774\ub860\uc801\uc778 \ubcf4\uc7a5\uc740 \uc5c6\uc2b5\ub2c8\ub2e4.<\/p>\n
from sklearn.model_selection import RandomizedSearchCV\nfrom scipy.stats import randint\n\n# Hyperparameter distribution\nparam_dist = {\n    'n_estimators': randint(10, 200),\n    'max_features': ['auto', 'sqrt', 'log2'],\n    'max_depth': [None] + list(range(10, 31)),\n}\n\nrandom_search = RandomizedSearchCV(estimator=RandomForestClassifier(), param_distributions=param_dist, n_iter=100, cv=3)\nrandom_search.fit(X_train, y_train)\nbest_params = random_search.best_params_\n<\/code><\/pre>\n
4. \ube44\ud2b8\ucf54\uc778 \uc790\ub3d9\ub9e4\ub9e4 \ubaa8\ub378 \uad6c\ucd95<\/h2>\n
\uc774\ubc88\uc5d0\ub294 \ube44\ud2b8\ucf54\uc778 \uac00\uaca9 \ub370\uc774\ud130\ub97c \uc218\uc9d1\ud558\uace0, \uc774\ub97c \ubc14\ud0d5\uc73c\ub85c \uc790\ub3d9\ub9e4\ub9e4\ub97c \uc704\ud55c \ub525\ub7ec\ub2dd \ubaa8\ub378\uc744 \uad6c\ucd95\ud558\uace0 \ud558\uc774\ud37c\ud30c\ub77c\ubbf8\ud130 \ud29c\ub2dd\uc758 \uc608\ub97c \uc0b4\ud3b4\ubcf4\uaca0\uc2b5\ub2c8\ub2e4.<\/p>\n
4.1 \ub370\uc774\ud130 \uc218\uc9d1<\/h3>\n
\ube44\ud2b8\ucf54\uc778 \uac00\uaca9 \ub370\uc774\ud130\ub294 \uc5ec\ub7ec \ub370\uc774\ud130 \uc81c\uacf5 \uc11c\ube44\uc2a4\uc5d0\uc11c API\ub97c \ud1b5\ud574 \uc218\uc9d1\ud560 \uc218 \uc788\uc2b5\ub2c8\ub2e4. \uc608\ub97c \ub4e4\uc5b4, Binance API\ub97c \ud1b5\ud574 \ub370\uc774\ud130\ub97c \uac00\uc838\uc62c \uc218 \uc788\uc2b5\ub2c8\ub2e4.<\/p>\n
import pandas as pd\nimport requests\n\ndef get_bitcoin_data():\n    url = 'https:\/\/api.binance.com\/api\/v3\/klines?symbol=BTCUSDT&interval=1d&limit=100'\n    response = requests.get(url)\n    data = response.json()\n    df = pd.DataFrame(data, columns=['Open time', 'Open', 'High', 'Low', 'Close', 'Volume', 'Close time', 'Quote asset volume', 'Number of trades', 'Taker buy base asset volume', 'Taker buy quote asset volume', 'Ignore'])\n    df['Close'] = df['Close'].astype(float)\n    df['Open time'] = pd.to_datetime(df['Open time'], unit='ms')\n    return df[['Open time', 'Close']]\n\nbitcoin_data = get_bitcoin_data()\n<\/code><\/pre>\n
4.2 \ub370\uc774\ud130 \uc804\ucc98\ub9ac<\/h3>\n
\uc218\uc9d1\ud55c \ub370\uc774\ud130\uc5d0 \ub300\ud574 \uc804\ucc98\ub9ac \uacfc\uc815\uc774 \ud544\uc694\ud569\ub2c8\ub2e4. \uacb0\uce21\uce58 \ucc98\ub9ac, \uc2a4\ucf00\uc77c\ub9c1, \uadf8\ub9ac\uace0 \ud6c8\ub828 \ub370\uc774\ud130\uc640 \ud14c\uc2a4\ud2b8 \ub370\uc774\ud130\ub85c \ub098\ub204\ub294 \uacfc\uc815\uc785\ub2c8\ub2e4.<\/p>\n
from sklearn.preprocessing import MinMaxScaler\nfrom sklearn.model_selection import train_test_split\n\n# \ub370\uc774\ud130 \uc804\ucc98\ub9ac\nscaler = MinMaxScaler()\nbitcoin_data['Close'] = scaler.fit_transform(bitcoin_data['Close'].values.reshape(-1, 1))\n\nX = bitcoin_data['Close'].shift(1).dropna().values.reshape(-1, 1)\ny = bitcoin_data['Close'].iloc[1:].values\n\nX_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)\n<\/code><\/pre>\n
4.3 \ubaa8\ub378 \uad6c\ucd95<\/h3>\n
LSTM(Long Short-Term Memory) \ub525\ub7ec\ub2dd \ubaa8\ub378\uc744 \uc0ac\uc6a9\ud558\uc5ec \ube44\ud2b8\ucf54\uc778 \uac00\uaca9 \uc608\uce21 \ubaa8\ub378\uc744 \uad6c\ucd95\ud569\ub2c8\ub2e4.<\/p>\n
from keras.models import Sequential\nfrom keras.layers import LSTM, Dense, Dropout\n\nmodel = Sequential()\nmodel.add(LSTM(units=50, return_sequences=True, input_shape=(1, 1)))\nmodel.add(Dropout(0.2))\nmodel.add(LSTM(units=50))\nmodel.add(Dropout(0.2))\nmodel.add(Dense(units=1))\n\nmodel.compile(optimizer='adam', loss='mean_squared_error')\n<\/code><\/pre>\n
4.4 \ubaa8\ub378 \ud559\uc2b5<\/h3>\n
\ubaa8\ub378\uc744 \ud559\uc2b5\uc2dc\ud0a4\uae30 \uc704\ud574 \ud558\uc774\ud37c\ud30c\ub77c\ubbf8\ud130 \uc870\uc815\uc774 \ud544\uc694\ud569\ub2c8\ub2e4. \ub2e4\uc74c\uc740 \ud559\uc2b5\ub960\uacfc \ubc30\uce58 \ud06c\uae30\ub97c \uc870\uc815\ud558\ub294 \uc608\uc81c\uc785\ub2c8\ub2e4.<\/p>\n
from keras.callbacks import EarlyStopping\n\nearly_stopping = EarlyStopping(monitor='loss', patience=3)\n\nmodel.fit(X_train.reshape((X_train.shape[0], 1, 1)), y_train, epochs=100, batch_size=1, callbacks=[early_stopping])\n<\/code><\/pre>\n
4.5 \uc608\uce21 \ubc0f \ud3c9\uac00<\/h3>\n
\ud559\uc2b5\ud55c \ubaa8\ub378\uc744 \ud1b5\ud574 \ud14c\uc2a4\ud2b8 \ub370\uc774\ud130\uc5d0 \ub300\ud55c \uc608\uce21\uc744 \uc218\ud589\ud558\uace0 \uc774\ub97c \ud3c9\uac00\ud569\ub2c8\ub2e4.<\/p>\n
import numpy as np\n\npredicted_prices = model.predict(X_test.reshape((X_test.shape[0], 1, 1)))\npredicted_prices = scaler.inverse_transform(predicted_prices)\n\n# \ubaa8\ub378 \ud3c9\uac00\nfrom sklearn.metrics import mean_squared_error\n\nmse = mean_squared_error(y_test, predicted_prices)\nprint('Mean Squared Error:', mse)\n<\/code><\/pre>\n
5. \uacb0\ub860<\/h2>\n
\uc774\ubc88 \uae00\uc5d0\uc11c\ub294 \ub525\ub7ec\ub2dd\uacfc \uba38\uc2e0\ub7ec\ub2dd\uc744 \ud65c\uc6a9\ud55c \ube44\ud2b8\ucf54\uc778 \uc790\ub3d9\ub9e4\ub9e4 \uc2dc\uc2a4\ud15c\uc758 \uad6c\ucd95 \uacfc\uc815\uacfc \ud568\uaed8, \ud558\uc774\ud37c\ud30c\ub77c\ubbf8\ud130 \ud29c\ub2dd\uc758 \uc911\uc694\uc131\uacfc \uadf8 \ubc29\ubc95\uc744 \uc790\uc138\ud788 \uc54c\uc544\ubcf4\uc558\uc2b5\ub2c8\ub2e4. \ud558\uc774\ud37c\ud30c\ub77c\ubbf8\ud130 \ud29c\ub2dd\uc744 \ud1b5\ud574 \ubaa8\ub378\uc758 \uc131\ub2a5\uc744 \ub04c\uc5b4\uc62c\ub9b4 \uc218 \uc788\uc73c\uba70, \uc774\ub97c \ud1b5\ud574 \ube44\ud2b8\ucf54\uc778 \uc790\ub3d9\ub9e4\ub9e4 \uc2dc\uc2a4\ud15c\uc758 \ud6a8\uc728\uc131\uc744 \ud06c\uac8c \ub192\uc77c \uc218 \uc788\uc2b5\ub2c8\ub2e4.<\/p>\n
6. \ucd94\uac00 \uc790\ub8cc<\/h3>\n
\ud558\uc774\ud37c\ud30c\ub77c\ubbf8\ud130 \ud29c\ub2dd\uc5d0 \uad00\ud55c \ub354 \ub9ce\uc740 \uc815\ubcf4\uc640 \uc790\ub8cc\ub294 \ub2e4\uc74c \ub9c1\ud06c\ub97c \ucc38\uc870\ud558\uc2dc\uae30 \ubc14\ub78d\ub2c8\ub2e4:<\/p>\n