gpt4 book ai didi

python - scikit-learn 中每个数据拆分的交叉验证指标

转载 作者:太空宇宙 更新时间:2023-11-04 08:29:21 25 4
gpt4 key购买 nike

我需要为 (X_test, y_test) 数据的每个拆分明确获取交叉验证统计信息。

所以,为了尝试这样做,我做了:

kf = KFold(n_splits=n_splits)

X_train_tmp = []
y_train_tmp = []
X_test_tmp = []
y_test_tmp = []
mae_train_cv_list = []
mae_test_cv_list = []

for train_index, test_index in kf.split(X_train):

for i in range(len(train_index)):
X_train_tmp.append(X_train[train_index[i]])
y_train_tmp.append(y_train[train_index[i]])

for i in range(len(test_index)):
X_test_tmp.append(X_train[test_index[i]])
y_test_tmp.append(y_train[test_index[i]])

model.fit(X_train_tmp, y_train_tmp) # FIT the model = SVR, NN, etc.

mae_train_cv_list.append( mean_absolute_error(y_train_tmp, model.predict(X_train_tmp)) # MAE of the train part of the KFold.

mae_test_cv_list.append( mean_absolute_error(y_test_tmp, model.predict(X_test_tmp)) ) # MAE of the test part of the KFold.

X_train_tmp = []
y_train_tmp = []
X_test_tmp = []
y_test_tmp = []

这是通过使用例如 KFold 获取每个交叉验证拆分的平均绝对误差 (MAE) 的正确方法吗?

最佳答案

您的方法存在一些问题。

首先,您当然不必在训练和验证列表中手动地逐一附加数据(即您的 2 个内部 for 循环);简单的索引就可以完成这项工作。

此外,我们通常从不计算和报告训练 CV 折叠的错误 - 仅计算验证折叠的错误。

牢记这些,并将术语切换为“验证”而不是“测试”,这是一个使用波士顿数据的简单可重现示例,它应该直接适应您的情况:

from sklearn.model_selection import KFold
from sklearn.datasets import load_boston
from sklearn.metrics import mean_absolute_error
from sklearn.tree import DecisionTreeRegressor

X, y = load_boston(return_X_y=True)
n_splits = 5
kf = KFold(n_splits=n_splits, shuffle=True)
model = DecisionTreeRegressor(criterion='mae')

cv_mae = []

for train_index, val_index in kf.split(X):
model.fit(X[train_index], y[train_index])
pred = model.predict(X[val_index])
err = mean_absolute_error(y[val_index], pred)
cv_mae.append(err)

之后,您的 cv_mae 应该是这样的(由于 CV 的随机性,细节会有所不同):

[3.5294117647058827,
3.3039603960396042,
3.5306930693069307,
2.6910891089108913,
3.0663366336633664]

当然,所有这些明确的东西并不是真正必要的;你可以用 cross_val_score 更简单地完成这项工作.不过有一个小问题:

from sklearn.model_selection import cross_val_score
cv_mae2 =cross_val_score(model, X, y, cv=n_splits, scoring="neg_mean_absolute_error")
cv_mae2
# result
array([-2.94019608, -3.71980198, -4.92673267, -4.5990099 , -4.22574257])

除了不是真正问题的负号外,您会注意到结果的方差与上面的 cv_mae 相比看起来明显更高;原因是我们没有洗牌我们的数据。不幸的是,cross_val_score 不提供混洗选项,因此我们必须使用 shuffle 手动执行此操作。所以我们最终的代码应该是:

from sklearn.model_selection import cross_val_score
from sklearn.utils import shuffle
X_s, y_s =shuffle(X, y)
cv_mae3 =cross_val_score(model, X_s, y_s, cv=n_splits, scoring="neg_mean_absolute_error")
cv_mae3
# result:
array([-3.24117647, -3.57029703, -3.10891089, -3.45940594, -2.78316832])

折叠之间的差异明显较小,并且更接近我们最初的 cv_mae...

关于python - scikit-learn 中每个数据拆分的交叉验证指标,我们在Stack Overflow上找到一个类似的问题: https://stackoverflow.com/questions/54201464/

25 4 0
Copyright 2021 - 2024 cfsdn All Rights Reserved 蜀ICP备2022000587号
广告合作:1813099741@qq.com 6ren.com