CODEWITHSUNDEEP
python
chad
chad

Reputation: 707

Translating equation into python

enter image description here

I am trying to compute this equation in python but I'm having problem understanding the loop and I am kinda confused at the summation of the summation.

This is how far i got and I have trouble implementing the summation of the summation. Any idea? 

import numpy as np
stdev =np.zeros(N)
m_sum=np.zeros([lam,N])

for i in xrange(N):
    for k in xrange(lam):
        m_sum[k,i]+=(1/(lam))*m[k,i]
    stdev[i]=1/(lam-1)*(m[k,i]-m_sum[k,i])**2

Upvotes: 0

Views: 954

Answers (1)

Amit Kumar Gupta
Amit Kumar Gupta

Reputation: 18617

It's confusing because the same index variable k is being used for both inner and outer loop, so it's ambiguous. However, since this is the standard formula for standard deviation, we know from context what the intention is:

import math
import numpy as np
stdev = np.zeros(N)
means = np.zeros(N)

for i in xrange(N):
    m = 0
    for k in xrange(lam):
        m += m[k,i]
    means[i] = m/lam

    v = 0
    for k in xrange(lam):
        v += (m[k,i] - means[i])**2
    stdev[i] = math.sqrt(v)/(lam-1)

Note that numpy also has functions to directly calculate standard deviations of columns of your matrix m, but if you're just doing this as an exercise of directly translating a math equation into code, then this is how you would do it.

Upvotes: 2

Related Questions