（十三）集成学习（下）——Stacking

参考：DataWhale教程链接

集成学习（上）所有Task：

（一）集成学习上——机器学习三大任务

（二）集成学习上——回归模型

（三）集成学习上——偏差与方差

（四）集成学习上——回归模型评估与超参数调优

（五）集成学习上——分类模型

（六）集成学习上——分类模型评估与超参数调优

（七）集成学习中——投票法

（八）集成学习中——bagging

（九）集成学习中——Boosting简介&AdaBoost

（十）集成学习中——GBDT

（十一）集成学习中——XgBoost、LightGBM

（十二）集成学习（下）——Blending

（十三）集成学习（下）——Stacking

Stacking集成学习算法

Stacking是一种比赛中常用的trick，严格它来说并不是一种算法，而是精美而又复杂的，对模型集成的一种策略。Stacking集成算法可以理解为一个两层的集成，第一层含有多个基础分类器，把输出的预测结果作为第二层的输入特征， 第二层的分类器通常是逻辑回归。

Blending存在的问题：Blending在第二层集成的时候中只会用了验证集的数据产生的特征，对数据的使用浪费比较大。

Stacking：采用交叉验证的思路，产生多组验证集，且可以充分利用训练集。

Blending与Stacking对比：

集成方法	Blending	Stacking
集成的特征	一次划分，特征简单，数据少	cv交叉验证，特征略复杂，数据多
泛化能力	可能会过拟合	健壮性好

看一下Stacking是如何集成算法的：(参考案例：https://www.cnblogs.com/Christina-Notebook/p/10063146.html)

由于sklearn并没有直接对Stacking的方法，因此我们需要下载mlxtend工具包(pip install mlxtend)

# 1. 简单堆叠3折CV分类
from sklearn import datasets

iris = datasets.load_iris()
X, y = iris.data[:, 1:3], iris.target
from sklearn.model_selection import cross_val_score
from sklearn.linear_model import LogisticRegression
from sklearn.neighbors import KNeighborsClassifier
from sklearn.naive_bayes import GaussianNB 
from sklearn.ensemble import RandomForestClassifier
from mlxtend.classifier import StackingCVClassifier

RANDOM_SEED = 42

clf1 = KNeighborsClassifier(n_neighbors=1)
clf2 = RandomForestClassifier(random_state=RANDOM_SEED)
clf3 = GaussianNB()
lr = LogisticRegression()

# Starting from v0.16.0, StackingCVRegressor supports
# `random_state` to get deterministic result.
sclf = StackingCVClassifier(classifiers=[clf1, clf2, clf3],  # 第一层分类器
                            meta_classifier=lr,   # 第二层分类器
                            random_state=RANDOM_SEED)

print('3-fold cross validation:\n')

for clf, label in zip([clf1, clf2, clf3, sclf], ['KNN', 'Random Forest', 'Naive Bayes','StackingClassifier']):
    scores = cross_val_score(clf, X, y, cv=3, scoring='accuracy')
    print("Accuracy: %0.2f (+/- %0.2f) [%s]" % (scores.mean(), scores.std(), label))

3-fold cross validation:

Accuracy: 0.91 (+/- 0.01) [KNN]
Accuracy: 0.95 (+/- 0.01) [Random Forest]
Accuracy: 0.91 (+/- 0.02) [Naive Bayes]
Accuracy: 0.93 (+/- 0.02) [StackingClassifier]

# 我们画出决策边界
from mlxtend.plotting import plot_decision_regions
import matplotlib.gridspec as gridspec
import itertools

gs = gridspec.GridSpec(2, 2)
fig = plt.figure(figsize=(10,8))
for clf, lab, grd in zip([clf1, clf2, clf3, sclf], 
                         ['KNN', 
                          'Random Forest', 
                          'Naive Bayes',
                          'StackingCVClassifier'],
                          itertools.product([0, 1], repeat=2)):
    clf.fit(X, y)
    ax = plt.subplot(gs[grd[0], grd[1]])
    fig = plot_decision_regions(X=X, y=y, clf=clf)
    plt.title(lab)
plt.show()

使用第一层所有基分类器所产生的类别概率值作为meta-classfier的输入。需要在StackingClassifier 中增加一个参数设置：use_probas = True。

另外，还有一个参数设置average_probas = True,那么这些基分类器所产出的概率值将按照列被平均，否则会拼接。

例如：

基分类器1：predictions=[0.2,0.2,0.7]

基分类器2：predictions=[0.4,0.3,0.8]

基分类器3：predictions=[0.1,0.4,0.6]

1）若use_probas = True，average_probas = True，

则产生的meta-feature 为：[0.233, 0.3, 0.7]

2）若use_probas = True，average_probas = False，

则产生的meta-feature 为：[0.2,0.2,0.7,0.4,0.3,0.8,0.1,0.4,0.6]

# 2.使用概率作为元特征
clf1 = KNeighborsClassifier(n_neighbors=1)
clf2 = RandomForestClassifier(random_state=1)
clf3 = GaussianNB()
lr = LogisticRegression()

sclf = StackingCVClassifier(classifiers=[clf1, clf2, clf3],
                            use_probas=True,  # 
                            meta_classifier=lr,
                            random_state=42)

print('3-fold cross validation:\n')

for clf, label in zip([clf1, clf2, clf3, sclf], 
                      ['KNN', 
                       'Random Forest', 
                       'Naive Bayes',
                       'StackingClassifier']):

    scores = cross_val_score(clf, X, y, 
                                              cv=3, scoring='accuracy')
    print("Accuracy: %0.2f (+/- %0.2f) [%s]" 
          % (scores.mean(), scores.std(), label))

3-fold cross validation:

Accuracy: 0.91 (+/- 0.01) [KNN]
Accuracy: 0.95 (+/- 0.01) [Random Forest]
Accuracy: 0.91 (+/- 0.02) [Naive Bayes]
Accuracy: 0.95 (+/- 0.02) [StackingClassifier]

# 3. 堆叠5折CV分类与网格搜索(结合网格搜索调参优化)
from sklearn.linear_model import LogisticRegression
from sklearn.neighbors import KNeighborsClassifier
from sklearn.naive_bayes import GaussianNB 
from sklearn.ensemble import RandomForestClassifier
from sklearn.model_selection import GridSearchCV
from mlxtend.classifier import StackingCVClassifier

# Initializing models

clf1 = KNeighborsClassifier(n_neighbors=1)
clf2 = RandomForestClassifier(random_state=RANDOM_SEED)
clf3 = GaussianNB()
lr = LogisticRegression()

sclf = StackingCVClassifier(classifiers=[clf1, clf2, clf3], 
                            meta_classifier=lr,
                            random_state=42)

params = {
     'kneighborsclassifier__n_neighbors': [1, 5],
          'randomforestclassifier__n_estimators': [10, 50],
          'meta_classifier__C': [0.1, 10.0]}

grid = GridSearchCV(estimator=sclf, 
                    param_grid=params, 
                    cv=5,
                    refit=True)
grid.fit(X, y)

cv_keys = ('mean_test_score', 'std_test_score', 'params')

for r, _ in enumerate(grid.cv_results_['mean_test_score']):
    print("%0.3f +/- %0.2f %r"
          % (grid.cv_results_[cv_keys[0]][r],
             grid.cv_results_[cv_keys[1]][r] / 2.0,
             grid.cv_results_[cv_keys[2]][r]))

print('Best parameters: %s' % grid.best_params_)
print('Accuracy: %.2f' % grid.best_score_)

0.947 +/- 0.03 {'kneighborsclassifier__n_neighbors': 1, 'meta_classifier__C': 0.1, 'randomforestclassifier__n_estimators': 10}
0.933 +/- 0.02 {'kneighborsclassifier__n_neighbors': 1, 'meta_classifier__C': 0.1, 'randomforestclassifier__n_estimators': 50}
0.940 +/- 0.02 {'kneighborsclassifier__n_neighbors': 1, 'meta_classifier__C': 10.0, 'randomforestclassifier__n_estimators': 10}
0.940 +/- 0.02 {'kneighborsclassifier__n_neighbors': 1, 'meta_classifier__C': 10.0, 'randomforestclassifier__n_estimators': 50}
0.953 +/- 0.02 {'kneighborsclassifier__n_neighbors': 5, 'meta_classifier__C': 0.1, 'randomforestclassifier__n_estimators': 10}
0.953 +/- 0.02 {'kneighborsclassifier__n_neighbors': 5, 'meta_classifier__C': 0.1, 'randomforestclassifier__n_estimators': 50}
0.953 +/- 0.02 {'kneighborsclassifier__n_neighbors': 5, 'meta_classifier__C': 10.0, 'randomforestclassifier__n_estimators': 10}
0.953 +/- 0.02 {'kneighborsclassifier__n_neighbors': 5, 'meta_classifier__C': 10.0, 'randomforestclassifier__n_estimators': 50}
Best parameters: {'kneighborsclassifier__n_neighbors': 5, 'meta_classifier__C': 0.1, 'randomforestclassifier__n_estimators': 10}
Accuracy: 0.95

# 如果我们打算多次使用回归算法，我们要做的就是在参数网格中添加一个附加的数字后缀，如下所示：
from sklearn.model_selection import GridSearchCV

# Initializing models

clf1 = KNeighborsClassifier(n_neighbors=1)
clf2 = RandomForestClassifier(random_state=RANDOM_SEED)
clf3 = GaussianNB()
lr = LogisticRegression()

sclf = StackingCVClassifier(classifiers=[clf1, clf1, clf2, clf3], 
                            meta_classifier=lr,
                            random_state=RANDOM_SEED)

params = {
     'kneighborsclassifier-1__n_neighbors': [1, 5],
          'kneighborsclassifier-2__n_neighbors': [1, 5],
          'randomforestclassifier__n_estimators': [10, 50],
          'meta_classifier__C': [0.1, 10.0]}

grid = GridSearchCV(estimator=sclf, 
                    param_grid=params, 
                    cv=5,
                    refit=True)
grid.fit(X, y)

cv_keys = ('mean_test_score', 'std_test_score', 'params')

for r, _ in enumerate(grid.cv_results_['mean_test_score']):
    print("%0.3f +/- %0.2f %r"
          % (grid.cv_results_[cv_keys[0]][r],
             grid.cv_results_[cv_keys[1]][r] / 2.0,
             grid.cv_results_[cv_keys[2]][r]))

print('Best parameters: %s' % grid.best_params_)
print('Accuracy: %.2f' % grid.best_score_)

0.940 +/- 0.02 {'kneighborsclassifier-1__n_neighbors': 1, 'kneighborsclassifier-2__n_neighbors': 1, 'meta_classifier__C': 0.1, 'randomforestclassifier__n_estimators': 10}
0.940 +/- 0.02 {'kneighborsclassifier-1__n_neighbors': 1, 'kneighborsclassifier-2__n_neighbors': 1, 'meta_classifier__C': 0.1, 'randomforestclassifier__n_estimators': 50}
0.940 +/- 0.02 {'kneighborsclassifier-1__n_neighbors': 1, 'kneighborsclassifier-2__n_neighbors': 1, 'meta_classifier__C': 10.0, 'randomforestclassifier__n_estimators': 10}
0.940 +/- 0.02 {'kneighborsclassifier-1__n_neighbors': 1, 'kneighborsclassifier-2__n_neighbors': 1, 'meta_classifier__C': 10.0, 'randomforestclassifier__n_estimators': 50}
0.960 +/- 0.02 {'kneighborsclassifier-1__n_neighbors': 1, 'kneighborsclassifier-2__n_neighbors': 5, 'meta_classifier__C': 0.1, 'randomforestclassifier__n_estimators': 10}
0.953 +/- 0.02 {'kneighborsclassifier-1__n_neighbors': 1, 'kneighborsclassifier-2__n_neighbors': 5, 'meta_classifier__C': 0.1, 'randomforestclassifier__n_estimators': 50}
0.953 +/- 0.02 {'kneighborsclassifier-1__n_neighbors': 1, 'kneighborsclassifier-2__n_neighbors': 5, 'meta_classifier__C': 10.0, 'randomforestclassifier__n_estimators': 10}
0.953 +/- 0.02 {'kneighborsclassifier-1__n_neighbors': 1, 'kneighborsclassifier-2__n_neighbors': 5, 'meta_classifier__C': 10.0, 'randomforestclassifier__n_estimators': 50}
0.960 +/- 0.02 {'kneighborsclassifier-1__n_neighbors': 5, 'kneighborsclassifier-2__n_neighbors': 1, 'meta_classifier__C': 0.1, 'randomforestclassifier__n_estimators': 10}
0.953 +/- 0.02 {'kneighborsclassifier-1__n_neighbors': 5, 'kneighborsclassifier-2__n_neighbors': 1, 'meta_classifier__C': 0.1, 'randomforestclassifier__n_estimators': 50}
0.953 +/- 0.02 {'kneighborsclassifier-1__n_neighbors': 5, 'kneighborsclassifier-2__n_neighbors': 1, 'meta_classifier__C': 10.0, 'randomforestclassifier__n_estimators': 10}
0.953 +/- 0.02 {'kneighborsclassifier-1__n_neighbors': 5, 'kneighborsclassifier-2__n_neighbors': 1, 'meta_classifier__C': 10.0, 'randomforestclassifier__n_estimators': 50}
0.953 +/- 0.02 {'kneighborsclassifier-1__n_neighbors': 5, 'kneighborsclassifier-2__n_neighbors': 5, 'meta_classifier__C': 0.1, 'randomforestclassifier__n_estimators': 10}
0.953 +/- 0.02 {'kneighborsclassifier-1__n_neighbors': 5, 'kneighborsclassifier-2__n_neighbors': 5, 'meta_classifier__C': 0.1, 'randomforestclassifier__n_estimators': 50}
0.953 +/- 0.02 {'kneighborsclassifier-1__n_neighbors': 5, 'kneighborsclassifier-2__n_neighbors': 5, 'meta_classifier__C': 10.0, 'randomforestclassifier__n_estimators': 10}
0.953 +/- 0.02 {'kneighborsclassifier-1__n_neighbors': 5, 'kneighborsclassifier-2__n_neighbors': 5, 'meta_classifier__C': 10.0, 'randomforestclassifier__n_estimators': 50}
Best parameters: {'kneighborsclassifier-1__n_neighbors': 1, 'kneighborsclassifier-2__n_neighbors': 5, 'meta_classifier__C': 0.1, 'randomforestclassifier__n_estimators': 10}
Accuracy: 0.96

# 4.在不同特征子集上运行的分类器的堆叠
##不同的1级分类器可以适合训练数据集中的不同特征子集。以下示例说明了如何使用scikit-learn管道和ColumnSelector：
from sklearn.datasets import load_iris
from mlxtend.classifier import StackingCVClassifier
from mlxtend.feature_selection import ColumnSelector
from sklearn.pipeline import make_pipeline
from sklearn.linear_model import LogisticRegression

iris = load_iris()
X = iris.data
y = iris.target

pipe1 = make_pipeline(ColumnSelector(cols=(0, 2)),  # 选择第0,2列
                      LogisticRegression())
pipe2 = make_pipeline(ColumnSelector(cols=(1, 2, 3)),  # 选择第1,2,3列
                      LogisticRegression())

sclf = StackingCVClassifier(classifiers=[pipe1, pipe2], 
                            meta_classifier=LogisticRegression(),
                            random_state=42)

sclf.fit(X, y)

StackingCVClassifier(classifiers=[Pipeline(steps=[('columnselector',
                                                   ColumnSelector(cols=(0, 2))),
                                                  ('logisticregression',
                                                   LogisticRegression())]),
                                  Pipeline(steps=[('columnselector',
                                                   ColumnSelector(cols=(1, 2,
                                                                        3))),
                                                  ('logisticregression',
                                                   LogisticRegression())])],
                     meta_classifier=LogisticRegression(), random_state=42)

# 5.ROC曲线 decision_function
### 像其他scikit-learn分类器一样，它StackingCVClassifier具有decision_function可用于绘制ROC曲线的方法。
### 请注意，decision_function期望并要求元分类器实现decision_function。
from sklearn import model_selection
from sklearn.linear_model import LogisticRegression
from sklearn.neighbors import KNeighborsClassifier
from sklearn.svm import SVC
from sklearn.ensemble import RandomForestClassifier
from mlxtend.classifier import StackingCVClassifier
from sklearn.metrics import roc_curve, auc
from sklearn.model_selection import train_test_split
from sklearn import datasets
from sklearn.preprocessing import label_binarize
from sklearn.multiclass import OneVsRestClassifier

iris = datasets.load_iris()
X, y = iris.data[:, [0, 1]], iris.target

# Binarize the output
y = label_binarize(y, classes=[0, 1, 2])
n_classes = y.shape[1]

RANDOM_SEED = 42

X_train, X_test, y_train, y_test = train_test_split(
    X, y, test_size=0.33, random_state=RANDOM_SEED)

clf1 =  LogisticRegression()
clf2 = RandomForestClassifier(random_state=RANDOM_SEED)
clf3 = SVC(random_state=RANDOM_SEED)
lr = LogisticRegression()

sclf = StackingCVClassifier(classifiers=[clf1, clf2, clf3],
                            meta_classifier=lr)

# Learn to predict each class against the other
classifier = OneVsRestClassifier(sclf)
y_score = classifier.fit(X_train, y_train).decision_function(X_test)

# Compute ROC curve and ROC area for each class
fpr = dict()
tpr = dict()
roc_auc = dict()
for i in range(n_classes):
    fpr[i], tpr[i], _ = roc_curve(y_test[:, i], y_score[:, i])
    roc_auc[i] = auc(fpr[i], tpr[i])

# Compute micro-average ROC curve and ROC area
fpr["micro"], tpr["micro"], _ = roc_curve(y_test.ravel(), y_score.ravel())
roc_auc["micro"] = auc(fpr["micro"], tpr["micro"])

plt.figure()
lw = 2
plt.plot(fpr[2], tpr[2], color='darkorange',
         lw=lw, label='ROC curve (area = %0.2f)' % roc_auc[2])
plt.plot([0, 1], [0, 1], color='navy', lw=lw, linestyle='--')
plt.xlim([0.0, 1.0])
plt.ylim([0.0, 1.05])
plt.xlabel('False Positive Rate')
plt.ylabel('True Positive Rate')
plt.title('Receiver operating characteristic example')
plt.legend(loc="lower right")
plt.show()