Numpy矩阵乘法误差

Question

我有两个大矩阵：

       Xn = np.matrix(X)
       Xnt = Xn.transpose()

然后Xn是这样的：

   >>> Xn
   matrix([['0,208', '0,22', '0,208', ..., '0,194', '0,205', '0,205'],
    ['0,22', '0,208', '0,214', ..., '0,205', '0,205', '0,214'],
    ['0,208', '0,214', '0,22', ..., '0,205', '0,214', '0,211'],
    ..., 
    ['0,214', '0,214', '0,208', ..., '0,199', '0,211', '0,226'],
    ['0,214', '0,208', '0,208', ..., '0,211', '0,226', '0,252'],
    ['0,208', '0,208', '0,211', ..., '0,226', '0,252', '0,24']], 
   dtype='|S5')

但我想要乘以Xn和Xnt。发生错误。

  >>> print(Xn*Xnt)

    Traceback (most recent call last):
    File "<pyshell#2>", line 1, in <module>
    print(Xn*Xnt)
    File "C:\Python27\lib\site-packages\numpy\matrixlib\defmatrix.py", line 330, in __mul__
    return N.dot(self, asmatrix(other))
    ValueError: data type must provide an itemsize

有什么问题？

Answer 1

您的矩阵元素是字符串 - 打印矩阵时请注意dtype='|S5'。尝试首先将X的元素从字符串转换为浮点数。