pandas Scikit-learn cross val score：数组的索引太多
声明：本页面是StackOverFlow热门问题的中英对照翻译，遵循CC BY-SA 4.0协议，如果您需要使用它，必须同样遵循CC BY-SA许可，注明原文地址和作者信息，同时你必须将它归于原作者（不是我）：StackOverFlow
原文地址: http://stackoverflow.com/questions/31995175/
Warning: these are provided under cc-by-sa 4.0 license. You are free to use/share it, But you must attribute it to the original authors (not me):
StackOverFlow
Scikit-learn cross val score: too many indices for array
提问by dartdog
I have the following code
我有以下代码
from sklearn.ensemble import ExtraTreesClassifier
from sklearn.cross_validation import cross_val_score
#split the dataset for train and test
combnum['is_train'] = np.random.uniform(0, 1, len(combnum)) <= .75
train, test = combnum[combnum['is_train']==True], combnum[combnum['is_train']==False]
et = ExtraTreesClassifier(n_estimators=200, max_depth=None, min_samples_split=10, random_state=0)
min_samples_split=10, random_state=0 )
labels = train[list(label_columns)].values
tlabels = test[list(label_columns)].values
features = train[list(columns)].values
tfeatures = test[list(columns)].values
et_score = cross_val_score(et, features, labels, n_jobs=-1)
print("{0} -> ET: {1})".format(label_columns, et_score))
Checking the shape of the arrays:
检查数组的形状：
features.shape
Out[19]:(43069, 34)
And
和
labels.shape
Out[20]:(43069, 1)
and I'm getting:
我得到：
IndexError: too many indices for array
and this relevant part of the traceback:
以及回溯的相关部分：
---> 22 et_score = cross_val_score(et, features, labels, n_jobs=-1)
I'm creating the data from Pandas dataframes and I searched here and saw some reference to possible errors via this method but can't figure out how to correct? What the data arrays look like: features
我正在从 Pandas 数据帧创建数据，我在这里搜索并看到了一些通过这种方法可能出现的错误的参考，但不知道如何纠正？数据数组的样子：特征
Out[21]:
array([[ 0., 1., 1., ..., 0., 0., 1.],
[ 0., 1., 1., ..., 0., 0., 1.],
[ 1., 1., 1., ..., 0., 0., 1.],
...,
[ 0., 0., 1., ..., 0., 0., 1.],
[ 0., 0., 1., ..., 0., 0., 1.],
[ 0., 0., 1., ..., 0., 0., 1.]])
labels
标签
Out[22]:
array([[1],
[1],
[1],
...,
[1],
[1],
[1]])
回答by YE LIANG HARRY
When we do cross validation in scikit-learn, the process requires an (R,)shape label instead of (R,1). Although they are the same thing to some extend, their indexing mechanisms are different. So in your case, just add:
当我们在 scikit-learn 中进行交叉验证时，该过程需要一个(R,)形状标签而不是(R,1)。尽管它们在某种程度上是相同的，但它们的索引机制是不同的。所以在你的情况下，只需添加：
c, r = labels.shape
labels = labels.reshape(c,)
before passing it to the cross-validation function.
在将其传递给交叉验证函数之前。
回答by Bud
It seems to be fixable if you specify the target labels as a single data column from Pandas. If the target has multiple columns, I get a similar error. For example try:
如果您将目标标签指定为 Pandas 的单个数据列，这似乎是可以修复的。如果目标有多个列，我会收到类似的错误。例如尝试：
labels = train['Y']
回答by MSalty
Adding .ravel()to the Y/Labels variable passed into the formula helped solve this problem within KNN as well.
添加.ravel()到传递给公式的 Y/Labels 变量也有助于解决 KNN 中的这个问题。
回答by Yang Zhao
try target:
尝试目标：
y=df['Survived']
instead , i used
相反，我用
y=df[['Survived']]
which made the target y a dateframe, it seems series would be ok
这使目标成为日期框架，看来系列没问题
回答by Gursel Karacor
You might need to play with the dimensions a bit, e.g.
您可能需要稍微调整一下尺寸，例如
et_score = cross_val_score(et, features, labels, n_jobs=-1)[:,n]
or
或者
et_score = cross_val_score(et, features, labels, n_jobs=-1)[n,:]
n being the dimension.
n 是维度。
