앙상블(Enemble)이란?

앙상블(Ensemble) 학습은 여러 개의 머신러닝 모델을 결합하여 단일 모델보다 더 나은 성능을 얻기 위한 기법을 말한다 앙상블 기법은 여러 개의 약한 학습기(weak learner)를 결합하여 강한 학습기(strong learner)를 생성하는 아이디어에 기반한다. 앙상블 기법은 다양한 머신러닝 문제에서 높은 성능을 달성하는 데 매우 효과적이다. 다양한 모델의 특징과 장점을 결합하므로, 오버피팅(과적합)을 줄이고 일반화 성능을 향상시킬 수 있다.

앙상블 학습의 주요 기법:

1. 배깅(Bagging, Bootstrap Aggregating):

   • 동일한 알고리즘에 대해 훈련 데이터의 서로 다른 부분 집합(subset)을 사용하여 여러 모델을 훈련한다.
   • 모든 모델의 예측을 집계하여 최종 예측을 생성한다.
   • ex) 랜덤 포레스트(Random Forest).

2. 부스팅(Boosting):

   • 연속적으로 모델을 훈련시키면서, 이전 모델의 오류를 다음 모델이 보정한다.
   • 모든 모델의 예측을 조합하여 최종 예측을 생성한다.
   • ex) AdaBoost, Gradient Boosting, XGBoost, LightGBM, CatBoost

3. 스태킹(Stacking):

   • 여러 다른 모델로부터의 예측을 취합하여, 그 예측들을 입력으로 사용하는 새로운 모델(메타 모델)을 훈련시킨다.
   • 이 메타 모델이 최종 예측을 생성한다.

---

배깅(Bagging) 예제 코드¶

Code View

Bootstrap Aggregating

    import numpy as np
    import matplotlib as mpl
    import matplotlib.pyplot as plt
    from sklearn.datasets import load_iris
    from sklearn.tree import DecisionTreeClassifier
    from sklearn.ensemble import BaggingClassifier

    iris = load_iris()
    X, y = iris.data[:, [0,2]], iris.target

    model1 = DecisionTreeClassifier(max_depth =10, random_state=0).fit(X, y)
    model2 = BaggingClassifier(DecisionTreeClassifier(max_depth=4), n_estimators=50, random_state=0).fit(X, y)

    x_min, x_max = X[:,0].min() - 1, X[:,0].max() + 1
    y_min, y_max = X[:,1].min() - 1, X[:,1].max() + 1
    xx, yy = np.meshgrid(np.arange(x_min, x_max, 0.1), np.arange(y_min, y_max, 0.1))

    plt.subplot(121)
    Z1 = model1.predict(np.c_[xx.ravel(), yy.ravel()]).reshape(xx.shape)
    plt.contourf(xx, yy, Z1, alpha=0.6, cmap=mpl.cm.jet)
    plt.scatter(X[:,0], X[:,1], c=y, alpha=1, s=50, cmap=mpl.cm.jet, edgecolors="k")
    plt.title("Decision tree")
    plt.subplot(122)

    Z2 = model2.predict(np.c_[xx.ravel(), yy.ravel()]).reshape(xx.shape)
    plt.contourf(xx, yy, Z2, alpha=0.6, cmap=mpl.cm.jet)
    plt.scatter(X[:,0], X[:,1],c=y,alpha=1,s=50,cmap=mpl.cm.jet,edgecolors="k")
    plt.title("Bagging of decision trees")
    plt.tight_layout()
    plt.show()

랜덤 포리스트 (Random Forest)

    import pandas as pd
    from sklearn import datasets
    from sklearn.model_selection import train_test_split
    from sklearn.ensemble import RandomForestClassifier
    from sklearn import metrics

    iris = datasets.load_iris()
    print('Class names :', iris.target_names)
    print('target : [0:setosa, 1:versicolor, 2:virginical]')
    print('No. of Data :', len(iris.data))
    print('Featrue names :', iris.feature_names)

    data = pd.DataFrame({
        'sepal length': iris.data[:,0], 'sepal width': iris.data[:,1], 'petal length': iris.data[:,2],
        'petal width':iris.data[:,3], 'species':iris.target
    })
    print(data.head()) # 일부 데이터 출력

    x = data[['sepal length', 'sepal width', 'petal length', 'petal width']] # 입력
    y = data['species'] # 출력
    x_train, x_test, y_train, y_test = train_test_split(x, y, test_size=0.3) # 테스트 데이터 30%
    print('No. of traing data: ', len(x_train))
    print('No. of test data:', len(y_test))

    forest = RandomForestClassifier(n_estimators=100) # 모델 생성
    forest.fit(x_train, y_train)

    y_pred = forest.predict(x_test) # 추론 (예측)
    print('Accuracy :', metrics.accuracy_score(y_test, y_pred))

Result

    Class names : ['setosa' 'versicolor' 'virginica']
    target : [0:setosa, 1:versicolor, 2:virginical]
    No. of Data : 150
    Featrue names : ['sepal length (cm)', 'sepal width (cm)', 'petal length (cm)', 'petal width (cm)']
    sepal length  sepal width  petal length  petal width  species
    0           5.1          3.5           1.4          0.2        0
    1           4.9          3.0           1.4          0.2        0
    2           4.7          3.2           1.3          0.2        0
    3           4.6          3.1           1.5          0.2        0
    4           5.0          3.6           1.4          0.2        0
    No. of traing data:  105
    No. of test data: 45
    Accuracy : 0.9333333333333333

배깅 회귀 (Bagging Regression)

    import numpy as np
    import pandas as pd
    from sklearn.datasets import load_boston # scikit-leanr < 1.2
    # from sklearn.datasets import fetch_california_housing # replace dataset
    from sklearn.metrics import mean_squared_error
    from sklearn.model_selection import train_test_split
    from sklearn.ensemble import BaggingRegressor
    from sklearn.tree import DecisionTreeRegressor
    import matplotlib.pyplot as plt

    boston = load_boston() # < 1.2
    data = pd.DataFrame(boston.data)
    data.columns = boston.feature_names
    data['PRICE'] = boston.target
    print(data.head())

    # replace dataset
    # california = fetch_california_housing()
    # data = pd.DataFrame(california.data)
    # data.columns = california.feature_names
    # data['PRICE'] = california.target
    # print(data.head())

    X, y = data.iloc[:,:-1],data.iloc[:,-1]
    X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=123)
    bag = BaggingRegressor(base_estimator = DecisionTreeRegressor( ), n_estimators = 10,
    max_features=1.0, bootstrap_features=False, random_state=0)
    bag.fit(X_train,y_train)
    preds = bag.predict(X_test)
    rmse = np.sqrt(mean_squared_error(y_test, preds))
    print("RMSE: %f" % (rmse))

Result

        CRIM    ZN  INDUS  CHAS    NOX     RM   AGE     DIS  RAD    TAX   
    0  0.00632  18.0   2.31   0.0  0.538  6.575  65.2  4.0900  1.0  296.0  
    1  0.02731   0.0   7.07   0.0  0.469  6.421  78.9  4.9671  2.0  242.0   
    2  0.02729   0.0   7.07   0.0  0.469  7.185  61.1  4.9671  2.0  242.0   
    3  0.03237   0.0   2.18   0.0  0.458  6.998  45.8  6.0622  3.0  222.0   
    4  0.06905   0.0   2.18   0.0  0.458  7.147  54.2  6.0622  3.0  222.0   

    PTRATIO       B  LSTAT  PRICE  
    0     15.3  396.90   4.98   24.0  
    1     17.8  396.90   9.14   21.6  
    2     17.8  392.83   4.03   34.7  
    3     18.7  394.63   2.94   33.4  
    4     18.7  396.90   5.33   36.2  
    RMSE: 4.594919

---

부스팅(Boosting) 예제 코드¶

Code View

AdaBoost - Regression

    import numpy as np
    import matplotlib.pyplot as plt
    from sklearn.tree import DecisionTreeRegressor
    from sklearn.ensemble import AdaBoostRegressor

    rng = np.random.RandomState(1)
    X = np.linspace(0, 6, 100)[:, np.newaxis]
    y = np.sin(X).ravel() + np.sin(6*X).ravel() + rng.normal(0, 0.1, X.shape[0])

    regr_1 = DecisionTreeRegressor(max_depth=4)
    regr_2 = AdaBoostRegressor(DecisionTreeRegressor(max_depth=4), n_estimators=100, random_state=rng)

    regr_1.fit(X, y)
    regr_2.fit(X, y)
    y_1 = regr_1.predict(X)
    y_2 = regr_2.predict(X)

    plt.figure()
    plt.scatter(X, y, c="k", label="training samples")
    plt.plot(X, y_1, c="g", label="n_estimators=1", linewidth=2)
    plt.plot(X, y_2, c="r", label="n_estimators=100", linewidth=2)
    plt.xlabel("data")
    plt.ylabel("target")
    plt.title("AdaBoost Regression")
    plt.legend()
    plt.show()

Gradient Boosting - Regression

import numpy as np
import pandas as pd
from sklearn import datasets
import matplotlib.pyplot as plt
from sklearn.model_selection import train_test_split
from sklearn.metrics import mean_squared_error
from sklearn import ensemble
from sklearn.metrics import mean_squared_error, r2_score
from sklearn.model_selection import cross_val_predict

boston = datasets.load_boston() # Boston 집값 데이터, 13개 속성, 마지막 중간값 정보
print(boston.data.shape, boston.target.shape)
print(boston.feature_names)

data = pd.DataFrame(boston.data, columns=boston.feature_names)
data = pd.concat([data, pd.Series(boston.target, name='MEDV')], axis=1)
print(data.head())
X = data.iloc[:,:-1]
y = data.iloc[:,-1]
x_training_set, x_test_set, y_training_set, y_test_set = train_test_split(X, y, test_size=0.10, random_state=42, shuffle=True)

Result

    (506, 13) (506,)
    ['CRIM' 'ZN' 'INDUS' 'CHAS' 'NOX' 'RM' 'AGE' 'DIS' 'RAD' 'TAX' 'PTRATIO'
    'B' 'LSTAT']
        CRIM    ZN  INDUS  CHAS    NOX     RM   AGE     DIS  RAD    TAX   
    0  0.00632  18.0   2.31   0.0  0.538  6.575  65.2  4.0900  1.0  296.0  
    1  0.02731   0.0   7.07   0.0  0.469  6.421  78.9  4.9671  2.0  242.0   
    2  0.02729   0.0   7.07   0.0  0.469  7.185  61.1  4.9671  2.0  242.0   
    3  0.03237   0.0   2.18   0.0  0.458  6.998  45.8  6.0622  3.0  222.0   
    4  0.06905   0.0   2.18   0.0  0.458  7.147  54.2  6.0622  3.0  222.0   

    PTRATIO       B  LSTAT  MEDV  
    0     15.3  396.90   4.98  24.0  
    1     17.8  396.90   9.14  21.6  
    2     17.8  392.83   4.03  34.7  
    3     18.7  394.63   2.94  33.4  
    4     18.7  396.90   5.33  36.2 

params = {'n_estimators':500, 'max_depth':4, 'min_samples_split':2, 'learning_rate':0.01, 'loss':'ls'}
model = ensemble.GradientBoostingRegressor(**params)
model.fit(x_training_set, y_training_set)
model_score = model.score(x_training_set, y_training_set)
print('R2 sq: ', model_score)

y_predicted = model.predict(x_test_set)
print('Mean squared error: %.2f'% mean_squared_error(y_test_set, y_predicted))
print('Test Variance score: %.2f' % r2_score(y_test_set, y_predicted))

fig, ax = plt.subplots()
ax.scatter(y_test_set, y_predicted, edgecolors=(0,0,0))
ax.plot([y_test_set.min(), y_test_set.max()], [y_test_set.min(), y_test_set.max()], 'k--', lw=4)
ax.set_xlabel('Actual')
ax.set_ylabel('Predicted')
ax.set_title('Ground Truth vs Predicted')
plt.show()

Gradient Boosting - Classification

    from sklearn.datasets import make_hastie_10_2
    from sklearn.ensemble import GradientBoostingClassifier
    import matplotlib.pyplot as plt

    X, y = make_hastie_10_2(random_state=0)
    X_train, X_test = X[:2000], X[2000:]
    y_train, y_test = y[:2000], y[2000:]
    print(X.shape, y.shape)
    print(X[0:5,:])
    print(y[0:5])

    clf = GradientBoostingClassifier(n_estimators=100, learning_rate=0.1, max_depth=1, random_state=0)
    clf.fit(X_train, y_train)
    print('Accuracy score (training): {0:.3f}'.format(clf.score(X_train, y_train)))
    print('Accuracy score (testing): {0:.3f}'.format(clf.score(X_test, y_test)))

Result

    (12000, 10) (12000,)
    [[ 1.76405235  0.40015721  0.97873798  2.2408932   1.86755799 -0.97727788
    0.95008842 -0.15135721 -0.10321885  0.4105985 ]
    [ 0.14404357  1.45427351  0.76103773  0.12167502  0.44386323  0.33367433
    1.49407907 -0.20515826  0.3130677  -0.85409574]
    [-2.55298982  0.6536186   0.8644362  -0.74216502  2.26975462 -1.45436567
    0.04575852 -0.18718385  1.53277921  1.46935877]
    [ 0.15494743  0.37816252 -0.88778575 -1.98079647 -0.34791215  0.15634897
    1.23029068  1.20237985 -0.38732682 -0.30230275]
    [-1.04855297 -1.42001794 -1.70627019  1.9507754  -0.50965218 -0.4380743
    -1.25279536  0.77749036 -1.61389785 -0.21274028]]
    [ 1. -1.  1. -1.  1.]
    Accuracy score (training): 0.879
    Accuracy score (testing): 0.819

XGBoosting - Regression

    import numpy as np
    import pandas as pd
    from sklearn.datasets import load_boston
    from sklearn.metrics import mean_squared_error
    from sklearn.model_selection import train_test_split
    import xgboost as xgb

    boston = load_boston()
    data = pd.DataFrame(boston.data)
    data.columns = boston.feature_names
    data['PRICE'] = boston.target
    print(data.head())
    X, y = data.iloc[:,:-1], data.iloc[:,-1]

    X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=123)
    xg_reg = xgb.XGBRegressor(objective='reg:squarederror', colsample_bytree=0.3, learning_rate=0.1, max_depth=5, alpha=10, n_estimators=10)
    xg_reg.fit(X_train, y_train)
    preds = xg_reg.predict(X_test)
    rmse = np.sqrt(mean_squared_error(y_test, preds))
    print('RMSE: %f' % (rmse))

Result

        CRIM    ZN  INDUS  CHAS    NOX     RM   AGE     DIS  RAD    TAX   
    0  0.00632  18.0   2.31   0.0  0.538  6.575  65.2  4.0900  1.0  296.0  
    1  0.02731   0.0   7.07   0.0  0.469  6.421  78.9  4.9671  2.0  242.0   
    2  0.02729   0.0   7.07   0.0  0.469  7.185  61.1  4.9671  2.0  242.0   
    3  0.03237   0.0   2.18   0.0  0.458  6.998  45.8  6.0622  3.0  222.0   
    4  0.06905   0.0   2.18   0.0  0.458  7.147  54.2  6.0622  3.0  222.0   

    PTRATIO       B  LSTAT  PRICE  
    0     15.3  396.90   4.98   24.0  
    1     17.8  396.90   9.14   21.6  
    2     17.8  392.83   4.03   34.7  
    3     18.7  394.63   2.94   33.4  
    4     18.7  396.90   5.33   36.2  
    RMSE: 10.423243

LightGBM

    from lightgbm import LGBMClassifier, LGBMRegressor
    from lightgbm import plot_importance, plot_metric, plot_tree
    from sklearn.datasets import load_iris
    from sklearn.model_selection import train_test_split
    from sklearn.model_selection import cross_validate

    iris = load_iris()
    X_train, X_test, y_train, y_test = train_test_split(iris.data, iris.target, test_size=0.2, random_state=123)
    lgbmc = LGBMClassifier(n_estimators=400)
    evals = [(X_test, y_test)]
    lgbmc.fit(X_train, y_train, early_stopping_rounds=100, eval_metric='logloss', eval_set=evals, verbose=True)
    preds = lgbmc.predict(X_test)

    cross_val = cross_validate(
        estimator=lgbmc,
        X=iris.data, y=iris.target,
        cv=5
    )

    print('avg fit time: {} (+/- {})'.format(cross_val['fit_time'].mean(), cross_val['fit_time'].std()))
    print('avg fit time: {} (+/- {})'.format(cross_val['score_time'].mean(), cross_val['score_time'].std()))
    print('avg fit time: {} (+/- {})'.format(cross_val['test_score'].mean(), cross_val['test_score'].std()))

    plot_metric(lgbmc)
    plot_importance(lgbmc, figsize=(10,12))
    plot_tree(lgbmc, figsize=(28,14))

Result

    [1]	valid_0's multi_logloss: 0.95847
    [2]	valid_0's multi_logloss: 0.832184
    [3]	valid_0's multi_logloss: 0.731164
    [4]	valid_0's multi_logloss: 0.641056
    [5]	valid_0's multi_logloss: 0.571726
    [6]	valid_0's multi_logloss: 0.507286
    [7]	valid_0's multi_logloss: 0.454933
    [8]	valid_0's multi_logloss: 0.410205
    [9]	valid_0's multi_logloss: 0.372194
    [10]	valid_0's multi_logloss: 0.333919
    [11]	valid_0's multi_logloss: 0.310212
    [12]	valid_0's multi_logloss: 0.282326
    [13]	valid_0's multi_logloss: 0.257165
    [14]	valid_0's multi_logloss: 0.240836
    [15]	valid_0's multi_logloss: 0.225383
    [16]	valid_0's multi_logloss: 0.211583
    [17]	valid_0's multi_logloss: 0.199289
    [18]	valid_0's multi_logloss: 0.186269
    [19]	valid_0's multi_logloss: 0.171556
    [20]	valid_0's multi_logloss: 0.168245
    [21]	valid_0's multi_logloss: 0.161065
    [22]	valid_0's multi_logloss: 0.151371
    [23]	valid_0's multi_logloss: 0.148081
    [24]	valid_0's multi_logloss: 0.143843
    [25]	valid_0's multi_logloss: 0.140169
    ...
    [137]	valid_0's multi_logloss: 0.376748
    avg fit time: 0.5514350891113281 (+/- 0.3701610138582717)
    avg fit time: 0.010002517700195312 (+/- 0.009552237668971902)
    avg fit time: 0.9600000000000002 (+/- 0.04898979485566355)