Python中的几种矩阵乘法
同线性代数中矩阵乘法的定义: np.dot()np.dot(A, B):对于二维矩阵,计算真正意义上的矩阵乘积,同线性代数中矩阵乘法的定义。对于一维矩阵,计算两者的内积。见如下Python代码:
import numpy as np
two_dim_matrix_one = np.array([[1, 2, 3], [4, 5, 6]])
two_dim_matrix_two = np.array([[1, 2], [3, 4], [5, 6]])
two_multi_res = np.dot(two_dim_matrix_one, two_dim_matrix_two) print(‘two_multi_res: %s’ %(two_multi_res))
one_dim_vec_one = np.array([1, 2, 3]) one_dim_vec_two = np.array([4, 5, 6]) one_result_res = np.dot(one_dim_vec_one, one_dim_vec_two) print(‘one_result_res: %s’ %(one_result_res))
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16结果如下:
two_multi_res: [[22 28] [49 64]] one_result_res: 32
1 2 3 对应元素相乘 element-wise product: np.multiply(), 或 *在Python中,实现对应元素相乘,有2种方式,一个是np.multiply(),另外一个是*。见如下Python代码:
import numpy as np
two_dim_matrix_one = np.array([[1, 2, 3], [4, 5, 6]]) another_two_dim_matrix_one = np.array([[7, 8, 9], [4, 7, 1]])
element_wise = two_dim_matrix_one * another_two_dim_matrix_one print(‘element wise product: %s’ %(element_wise))
element_wise_2 = np.multiply(two_dim_matrix_one, another_two_dim_matrix_one) print(‘element wise product: %s’ % (element_wise_2))
1 2 3 4 5 6 7 8 9 10 11 12 13结果如下:
element wise product: [[ 7 16 27] [16 35 6]] element wise product: [[ 7 16 27] [16 35 6]]