python - sklearn.model_selection.train_test_split中的样本权重是什么

Question

在 probability calibration of classifiers来自scikit-learn，有一段关于train_test_split的代码我在文档中找不到解释。

centers = [(-5, -5), (0, 0), (5, 5)] X, y =
make_blobs(n_samples=n_samples, n_features=2, cluster_std=1.0,
                  centers=centers, shuffle=False, random_state=42)

y[:n_samples // 2] = 0 
y[n_samples // 2:] = 1 
sample_weight = np.random.RandomState(42).rand(y.shape[0])

# split train, test for calibration 
X_train, X_test, y_train, y_test, sw_train, sw_test = \
    train_test_split(X, y, sample_weight, test_size=0.9, random_state=42)

1. What does sample_weight in train_test_split do?

2. How does the source code of train_test_split process sample_weight?

提前非常感谢。

Answer 1

train_test_split 不仅仅采用 x 和 y。它可以采用具有相同第一维度的任意数组序列，并将它们随机但一致地分成沿该维度的两组。

在您的示例中，有一个随机权重数组(每个观察一个权重)，该数组被分为训练数组和测试数组:sw_train 和 sw_test。

为观测值分配权重的原因有很多。如需进一步讨论，请参阅:

• https://blogs.sas.com/content/iml/2017/10/02/weight-variables-in-statistics-sas.html
• https://www.r-bloggers.com/linear-models-with-weighted-observations/
• http://www.applied-survey-methods.com/weight.html