有没有办法解决这个关于 StratifiedShuffleSplit 的错误？

问题描述

我是 ML 的新手，我一直在尝试 udacity ML 项目。但是，我遇到了一个错误，我很难解决。代码看起来没问题，但我似乎无法遍历数据。我知道这与所做的新 StratifiedShuffleSplit 更改有关。代码已关闭。

def Stratified_Shuffle_Split(X,y,num_test):
    sss = StratifiedShuffleSplit(y,1,test_size=num_test,random_state = None)
    for train,test in sss:
        X_train,X_test = X.iloc[train],X.iloc[test]
        y_train,y_test = y.iloc[train],y.iloc[test]
    return X_train,X_test,y_train,y_test

# First,decide how many training vs test samples you want
num_all = student_data.shape[0]  # same as len(student_data)
num_train = round(num_all*0.75)  # about 75% of the data
num_test = num_all - num_train
#print(num_test)

y = student_data['passed'] # identify target variable
X_train,y_test = Stratified_Shuffle_Split(X_all,num_test)

print("Training Set: {0:.2f} Samples".format(X_train.shape[0]))
print("Testing Set: {0:.2f} Samples".format(X_test.shape[0]))

我的错误是这个

---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
<ipython-input-20-2147158fcaf2> in <module>
     13 
     14 y = student_data['passed'] # identify target variable
---> 15 X_train,num_test)
     16 
     17 print("Training Set: {0:.2f} Samples".format(X_train.shape[0]))

<ipython-input-20-2147158fcaf2> in Stratified_Shuffle_Split(X,num_test)
      1 def Stratified_Shuffle_Split(X,num_test):
      2     sss = StratifiedShuffleSplit(y,random_state = None)
----> 3     for train,test in sss:
      4         X_train,X.iloc[test]
      5         y_train,y.iloc[test]

TypeError: 'StratifiedShuffleSplit' object is not iterable
'''

解决方法

根据documentation，需要在StratifiedShuffleSplit上运行.split()函数。您需要 .split() 来生成要切片的索引。所以这部分可能是：

def Stratified_Shuffle_Split(X,y,num_test):
     sss = StratifiedShuffleSplit(y,1,test_size=num_test,random_state = None)
     for train,test in sss.split(X,y):
        X_train,X_test = X.iloc[train],X.iloc[test]
        y_train,y_test = y.iloc[train],y.iloc[test]
     return X_train,X_test,y_train,y_test

我也不确定是否需要定义一个新函数，StratifiedShuffleSplit 已经是一个现成的函数，可以使用您拥有的 for 循环执行您想要的操作。