交叉验证样本数不一致错误Python

如何解决交叉验证样本数不一致错误Python

我正在尝试使用交叉验证方法和SVM分类器进行分类。在我的数据文件中，最后一列包含我的类（分别是0、1、2、3、4、5），其余（第一列除外）是我要用来预测这些类的数字数据。>

from sklearn import svm
from sklearn import metrics
import numpy as np
from sklearn.model_selection import StratifiedKFold
from sklearn.model_selection import cross_val_score


filename = "Features.csv"
dataset = np.loadtxt(filename,delimiter=',',skiprows=1,usecols=range(1,39))

x = dataset[:,0:36]
y = dataset[:,36]
print("len(x): " + str(len(x)))
print("len(y): " + str(len(x)))

skf = StratifiedKFold(n_splits=10,shuffle=False,random_state=42)

modelsvm = svm.SVC()

expected = y
print("len(expected): " + str(len(expected)))

predictedsvm = cross_val_score(modelsvm,x,y,cv=skf)
print("len(predictedsvm): " + str(len(predictedsvm)))

svm_results = metrics.classification_report(expected,predictedsvm)

print(svm_results)

我收到这样的错误：

len(x): 2069
len(y): 2069
len(expected): 2069
C:\Python\Python37\lib\site-packages\sklearn\model_selection\_split.py:297: FutureWarning: Setting a random_state has no effect since shuffle is False. This will raise an error in 0.24. You should leave random_state to its default (None),or set shuffle=True.
  FutureWarning
len(predictedsvm): 10
Traceback (most recent call last):
  File "C:/Users/MyComp/PycharmProjects/GG/AR.py",line 54,in <module>
    svm_results = metrics.classification_report(expected,predictedsvm)
  File "C:\Python\Python37\lib\site-packages\sklearn\utils\validation.py",line 73,in inner_f
    return f(**kwargs)
  File "C:\Python\Python37\lib\site-packages\sklearn\metrics\_classification.py",line 1929,in classification_report
    y_type,y_true,y_pred = _check_targets(y_true,y_pred)
  File "C:\Python\Python37\lib\site-packages\sklearn\metrics\_classification.py",line 81,in _check_targets
    check_consistent_length(y_true,y_pred)
  File "C:\Python\Python37\lib\site-packages\sklearn\utils\validation.py",line 257,in check_consistent_length
    " samples: %r" % [int(l) for l in lengths])
ValueError: Found input variables with inconsistent numbers of samples: [2069,10]

Process finished with exit code 1

当我尝试使用CV进行预测时，我不知道y中的数据如何减少到10。

有人可以帮我吗？

解决方法

您误解了cross_val_score的输出。按照documentation，它返回“每次交叉验证的估算器分数数组”，而不是实际的预测。因为有10折，所以得到10个值。

classification_report期望真实值和预测值。要使用此功能，您需要使用模型进行预测。为此，您需要使模型适合数据。如果您对cross_val_score的结果感到满意，则可以在数据上训练该模型。或者，您可以使用GridSearchCV一次完成所有操作。