Python 中一维矩阵的成对计算

Question

My question is basically the same as this post here ...and maybe it's exactly the answer I just don't understand the syntax of the np.arrange() function.

Here is what I am trying to do:

For each unique pair of values in a 1D array I want to calculate the ratio of max value to min value. All the examples I can find, including the above mentioned, would indicate the np.arrange() function only works with sequential values ie [1,2,3,4,5] and not from a unique array such as [1.34, 87.5, 2.10, 700.4]

max(0,1)/min(0,1) = (87.5/1.34)

max(0,2)/min(0,2) = (2.10/1.34)

Is the np.arrange() what I should be using or is there another function better suited to solving the problem?

Answer 1

np.arange() has not been used to solve the problem, rather numpy broadcasting has been used to provide a solution in the link shared by you. np.arange() is used to create the initial array to work on (see doc ).

A simple solution to your problem would be using 2 loops like:

x = [1.34, 87.5, 2.10, 700.4]
mat = np.zeros(shape=(len(x),len(x)))
for i, ival in enumerate(x):
  for j, jval in enumerate(x):
    mat[i][j]=max(ival,jval)/min(ival,jval)

This yields the pairwise ratio of max to min:

 Result:
        [[  1.     65.299   1.567 522.687]
         [ 65.299   1.     41.667   8.005]
         [  1.567  41.667   1.    333.524]
         [522.687   8.005 333.524   1.   ]]