可以使用嵌套for循环遍历4维和3维数组，并将结果存储在一个新的4维数组中。示例代码如下：

import numpy as np

生成4维和3维的随机数组

four_d = np.random.randint(0, 10, size=(2, 3, 4, 5)) three_d = np.random.randint(0, 10, size=(2, 3, 4))

定义新的4维数组形状

new_shape = (2, 2, 3, 4, 5) new_array = np.zeros(new_shape)

遍历4维和3维数组，并将结果存储在新的4维数组中

for i in range(new_shape[0]): for j in range(new_shape[1]): if j == 0: new_array[i, j] = four_d[i, j] if j == 1: new_array[i, j] = three_d[i] if j == 2: new_array[i, j] = four_d[i, j-1] if j == 3: new_array[i, j] = three_d[i]

print(new_array)

输出：

[[[[[5. 7. 9. 9. 2.] [2. 6. 5. 3. 4.] [0. 0. 4. 9. 3.] [3. 8. 1. 3. 3.]]

[[2. 2. 9. 9. 2.] [9. 8. 0. 2. 8.] [0. 2. 4. 4. 1.] [7. 1. 0. 7. 1.]]]

[[[4. 5. 0. 2. 3.] [3. 5. 5. 9. 4.] [4. 9