Python generate multiple random matrices in a for loop

Question

I am trying to generate 100 random column stochastic matrices using a for loop and save them.

Below code is to generate an individual random column stochastic matrix.

import numpy as np
import random

For x in range(100):
    a = 2
    sto = np.identity(2)
    sto = sto + np.random.uniform(low=0.2, high=0.5, size=(a, a))
    sto = sto / sto.sum(axis=0, keepdims=1)
    print(sto)

~~When I try to implement the operation to generate 100 matrices with a for loop I fail - all matrices generated turn out to be the same.~~

Edit: I can now generate different matrices in the loop.

But I still don't know how to save them all to use in later work (for instance: saving them to multiply some of them at a later step)?

Sorry, I just edited it. I think I solved the problem with generating the same matrix over and over again: the ones generated now are different from each other. But I still have the problem with saving them. — experimental music
– experimental music, Commented Jun 30, 2019 at 0:57

Mehdi · Accepted Answer · 2019-06-30 02:05:18Z

1

I'm not sure if this is what you are looking for, but basically i created a list named sto1 so that you can access it in the later parts of your code

import numpy as np
import random

sto1 = []
for x in range(100):
    a = 2
    sto = np.identity(2)
    sto = sto + np.random.uniform(low=0.2, high=0.5, size=(a, a))
    sto = sto / sto.sum(axis=0, keepdims=1)
    sto1.append(sto)
print(sto1)

edited Jun 30, 2019 at 2:05

Mehdi

4,3385 gold badges24 silver badges36 bronze badges

answered Jun 30, 2019 at 1:44

Axois

2,0612 gold badges13 silver badges23 bronze badges

Sign up to request clarification or add additional context in comments.

Comments

Mehdi · Accepted Answer · 2019-06-30 02:04:19Z

0

I am not sure exactly what you expect as the outcome, but I don't think you need a for-loop.

Here is my solution:

a = 2
i = np.expand_dims(np.eye(a), 0)
x = np.random.uniform(low=0.2, high=0.5, size=(100, a, a))
s = i + x
out = s / np.expand_dims(s.sum(1), 1)

In fact, internally a 100-loop adds all values in matching dimensions of i and x. The trick is that expand_dims method creates a dummy dimension to be able to only have one mismatched dimension equal to 1. Then, these operations between two arrays are possible. The second expand_dims also has a similar role but on a different dimension. In the end, if you want it as a list, you can convert it to a list of 100 arrays:

out = list(out)

edited Jun 30, 2019 at 2:04

answered Jun 30, 2019 at 1:48

Mehdi

4,3385 gold badges24 silver badges36 bronze badges

Comments

some_name.py · Accepted Answer · 2019-07-01 07:42:18Z

0

Why not just save them on the hard-drive:

import numpy as np

random_matrices = np.random.rand(100,2,2)

np.save('random_matrices.npy', random_matrices)

random_matrices_loaded = np.load('random_matrices.npy')

answered Jul 1, 2019 at 7:42

some_name.py

8377 silver badges17 bronze badges

Collectives™ on Stack Overflow

Python generate multiple random matrices in a for loop

3 Answers 3

Comments

Comments

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

3 Answers 3

Comments

Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Related