使用矩阵作为字典值并在python中执行numpy矩阵运算

Question

I have a dictionary with key and value (it was given to me). When I read a key to extract the value, I get something like this: matrix([[1.234, -4.056]])

I call this value A.

I define a variable as below

B = np.matrix([0, 0])

B is a running sum of A times a group of scalars,

for i in range(0, n):

     B =+ A*scalar[i]

The problem is that the output has the format

matrix([[xxx , yyy]])

and I need

matrix([xxx, yyy])

that is, I do not want the double brackets.

Answer 1

You want a numpy.array not a numpy.matrix . The np.matrix docs is a 2D data structure, np.array is an nd dimensional structure.

If you look at B.shape immediately after creating you will discover that it is (1,2) , not (2,) as you intended.

B.A.reshape(2,)  # or B.A1

will give you a np.array that is 1-dimensional.

NumPy arrays ARE NOT equivalent to NumPy matrices, the meaning of operations is different between the two, for instance * is a dot product for matrix but an element wise product for an array .