构建函数 — 划分测试集和训练集

import numpy as np
def split_train_test(data, test_ratio):
	shuffled_indices = np.random.permutation(len(data))
	test_set_size = int(len(data) * test_ratio)
	test_indices = shuffled_indices[:test_set_size]
	train_indices = shuffled_indices[test_set_size:]
	return data.iloc[train_indices], data.iloc[test_indices]

train_set, test_set = split_train_test(housing, 0.2)
print(len(train_set), "train +", len(test_set), "test")