C:\Users\pcl>conda activate torch38
(torch38) C:\Users\pcl>python
Python 3.8.11 (default, Aug 6 2021, 09:57:55) [MSC v.1916 64 bit (AMD64)] :: Anaconda, Inc. on win32
Type "help", "copyright", "credits" or "license" for more information.
>>> from sklearn.model_selection import KFold
>>> import numpy as np
>>> import pandas as pd
>>> xt=np.random.randint(5,27,20)
>>> xt = xt.reshape(5,4)
>>> xt
array([[17, 13, 23, 26],
[ 8, 10, 13, 8],
[24, 7, 10, 23],
[21, 13, 13, 21],
[10, 9, 12, 10]])
>>> yt = np.random.randint(0,3,5)
>>> yt
array([1, 1, 1, 0, 2])
>>> folds = KFold(n_splits=5, shuffle = True, random_state=2021)
>>> xt = pd.DataFrame(xt)
>>> for tidx , vidx in folds.split(xt, yt):
... vdf,vl = xt.iloc[vidx,:], yt[vidx]
... tdf, tl = xt.iloc[tidx,:], yt[tidx]
... print('fold id',tidx, vidx)
...
fold id [0 1 3 4] [2]
fold id [1 2 3 4] [0]
fold id [0 1 2 4] [3]
fold id [0 2 3 4] [1]
fold id [0 1 2 3] [4]
>>>


这段代码演示了如何在Python中使用scikit-learn库的KFold类进行交叉验证。首先,它生成了一些随机数据,然后用这些数据创建了一个DataFrame。接着,它设置了KFold的参数并进行了拆分,输出了每次拆分的训练集和验证集的索引。
1490

被折叠的 条评论
为什么被折叠?



