Yes, numbers of the order e-16 can be considered machine-precision level (for numpy with basic settings) and thus equal to zero.
Note that, before normalizing, the mean(s) should be non-zero, because np.random.rand will yield random numbers uniformly in the interval [0,1] (Can you guess what the expected mean here is?)
Lab/ex01 task A
Hello,
Is it possible that some rounding errors lead to have a very small mean (e-16 order of magnitude)?
Yes, numbers of the order e-16 can be considered machine-precision level (for numpy with basic settings) and thus equal to zero.
Note that, before normalizing, the mean(s) should be non-zero, because
np.random.randwill yield random numbers uniformly in the interval[0,1](Can you guess what the expected mean here is?)Thak you :)
I guess the expected mean will be 0.5 given the np.random.rand distribution
1