深度学习--十折交叉验证

用scikit-learn来评价模型质量，为了更好地挑拣出结果的差异，采用了十折交叉验证（10-fold cross validation）方法。

本程序在输入层和第一个隐含层之间加入20%Dropout

采用十折交叉验证的方法进行测试。

    # dropout in the input layer with weight constraint
    def create_model1():
        # create model
        model = Sequential()
        model.add(Dropout(0.2, input_shape=(60,)))
        model.add(Dense(60, init='normal', activation='relu', W_constraint=maxnorm(3)))
        model.add(Dense(30, init='normal', activation='relu', W_constraint=maxnorm(3)))
        model.add(Dense(1, init='normal', activation='sigmoid'))
        # Compile model
        sgd = SGD(lr=0.1, momentum=0.9, decay=0.0, nesterov=False)
        model.compile(loss='binary_crossentropy', optimizer=sgd, metrics=['accuracy'])
        return model
 
    numpy.random.seed(seed)
    estimators = []
    estimators.append(('standardize', StandardScaler()))
    estimators.append(('mlp', KerasClassifier(build_fn=create_model1, nb_epoch=300, batch_size=16, verbose=0)))
    pipeline = Pipeline(estimators)
    kfold = StratifiedKFold(y=encoded_Y, n_folds=10, shuffle=True, random_state=seed)
    results = cross_val_score(pipeline, X, encoded_Y, cv=kfold)
    print("Accuracy: %.2f%% (%.2f%%)" % (results.mean()*100, results.std()*100))

Pineline

from sklearn.pipeline import Pipeline
from sklearn.preprocessing import StandardScaler
num_pipeline = Pipeline([
	('imputer', Imputer(strategy="median")),
	('attribs_adder', CombinedAttributesAdder()),
	('std_scaler', StandardScaler()),
])
housing_num_tr = num_pipeline.fit_transform(housing_num)

Pipeline构造器接受(name, transform) tuple的列表作为参数。按顺序执行列表中的transform，完成数据预处理

StratifiedKFold

StratifiedKFold用法类似Kfold，但是分层采样，确保训练集，测试集中各类别样本的比例与原始数据集中相同

sklearn.model_selection.StratifiedKFold(n_splits=3, shuffle=False, random_state=None) 

import numpy as np 
from sklearn.model_selection import KFold,StratifiedKFold
 
X=np.array([
    [1,2,3,4],
    [11,12,13,14],
    [21,22,23,24],
    [31,32,33,34],
    [41,42,43,44],
    [51,52,53,54],
    [61,62,63,64],
    [71,72,73,74]
])
 
y=np.array([1,1,0,0,1,1,0,0])
floder = KFold(n_splits=4,random_state=0,shuffle=False)
sfolder = StratifiedKFold(n_splits=4,random_state=0,shuffle=False)
 
for train, test in sfolder.split(X,y):
    print('Train: %s | test: %s' % (train, test))
    print(" ")
 
for train, test in floder.split(X,y):
    print('Train: %s | test: %s' % (train, test))

#RESULT
Train: [1 3 4 5 6 7] | test: [0 2]
 
Train: [0 2 4 5 6 7] | test: [1 3]
 
Train: [0 1 2 3 5 7] | test: [4 6]
 
Train: [0 1 2 3 4 6] | test: [5 7]
 
Train: [2 3 4 5 6 7] | test: [0 1]
 
Train: [0 1 4 5 6 7] | test: [2 3]
 
Train: [0 1 2 3 6 7] | test: [4 5]
 
Train: [0 1 2 3 4 5] | test: [6 7]

cross_val_score：

不同的训练集、测试集分割的方法导致其准确率不同
交叉验证的基本思想是：将数据集进行一系列分割，生成一组不同的训练测试集，然后分别训练模型并计算测试准确率，最后对结果进行平均处理。这样来有效降低测试准确率的差异。

使用交叉验证的建议

1. K=10是一个一般的建议

2. 如果对于分类问题，应该使用分层抽样（stratified sampling）来生成数据，保证正负例的比例在训练集和测试集中的比例相同

from sklearn.cross_validation import cross_val_score

knn = KNeighborsClassifier(n_neighbors=5)
# 这里的cross_val_score将交叉验证的整个过程连接起来，不用再进行手动的分割数据
# cv参数用于规定将原始数据分成多少份
scores = cross_val_score(knn, X, y, cv=10, scoring='accuracy')
print(scores)
print(scores.mean())#输出结果平均值

参考网页：

https://blog.csdn.net/u010159842/article/details/54138157

cross_val_score交叉验证及其用于参数选择、模型选择、特征选择

https://blog.csdn.net/u012735708/article/details/82258615