Strange behavior in generating Numpy array with random values

Question

With this code I want to generate some small numbers in Numpy array format:

np.random.seed(1)
syn0 = 2*np.random.random((315,1))-1

this gives me result like this:

[[-0.16595599]
 [ 0.44064899]
 [-0.99977125]
 [-0.39533485]
 [ 0.70648822]
 [-0.81532281]
 [-0.62747958]
 [ 0.30887855]
 [-0.20646505]
 [ 0.07763347]
 [-0.16161097]
      .
      .
      .

But when i change rows count to 316 and upper:

np.random.seed(1)
syn0 = 2*np.random.random((316,1))-1

Then I get results in this number format(power notation):

[[ -1.65955991e-01]
 [  4.40648987e-01]
 [ -9.99771250e-01]
 [ -3.95334855e-01]
 [ -7.06488218e-01]
 [ -8.15322810e-01]
 [ -6.27479577e-01]
 [ -3.08878546e-01]
 [ -2.06465052e-01]
 [  7.76334680e-02]
 [ -1.61610971e-01]
 [  3.70439001e-01]
         .
         .
         .

I know this is power notation format but why this happens? I dont need this format.Why this strange behavior happens?

It's just an representation. Nothing within your array changes. Why this is happening? Probably because the spread of values is likely to increase when taking more samples and there is some threshold within numpy's print function. Of course you can tell numpy to print it as you like. Consider reading the docs for this. — sascha
– sascha, Commented Nov 22, 2016 at 22:03

donkopotamus · Accepted Answer · 2016-11-22 22:32:56Z

2

You can print a particular array without scientific notation using numpy.array2string:

print(np.array2string(x, suppress_small=True))

To not use scientific notation by default for all printing, use numpy.set_printoptions:

np.set_printoptions(suppress=True)

As to why this happens - when printing a floating point array the default behaviour is to use scientific notation if:

the minimum (absolute non zero) value to be printed is less than 0.0001; or
the ratio of the maximum (absolute non zero) value to minimum (absolute non zero) value is greater than 1000.

This behaviour is suppressed if suppress_small is passed to array2string (or suppress to set_printoptions)

edited Nov 22, 2016 at 22:32

answered Nov 22, 2016 at 22:26

donkopotamus

23.4k3 gold badges58 silver badges61 bronze badges

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

Comments

timleathart · Accepted Answer · 2016-11-22 22:10:36Z

1

You can turn off scientific notation by changing numpy's print options:

np.set_printoptions(suppress=True)

syn0
>>> array([[-0.00083166],
           [ 0.45717134],
           [-0.58361112],
           [-0.50393288],
           [ 0.70334375],
           [-0.16830256],
           [ 0.23337013],
           [-0.53266772],
              ....

answered Nov 22, 2016 at 22:10

timleathart

5801 gold badge6 silver badges21 bronze badges

Collectives™ on Stack Overflow

Strange behavior in generating Numpy array with random values

2 Answers 2

Comments

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Related