Calculate entropy of a lists of probabilities in python-CodePudding

I have a dictionary like this:

a = {'a': [0.2,0.3,0.6], 'b': [0.4,0.5,0.9], 'c': [0.7,0.1,0.6]}

I want to calculate the overall entropy of each list of this dictionary. To this end, I first need to calculate the entropy for each of probabilities within each list. I tried the below code:

lista_all = []
for i in a.values():
  c = [0 - (p * -log2(p)) for p in i ]
  lista_all.append(c)

But my result show that the loop is appending the same list to my empty list. So, my current result is:

[[-0.46438561897747244, -0.5210896782498619, 
-0.44217935649972373], [-0.46438561897747244, -0.5210896782498619, 
-0.44217935649972373], [-0.46438561897747244, -0.5210896782498619, 
-0.44217935649972373]]

My desired output would be something like:

[
 [-0.464, -0.521,-0.442], 
 [-0.529, -0.500, --0.137], 
 [-0.360, -0.332, -0.521]
]

Can someone help me to achieve this goal?

CodePudding user response：

Adding the round(num, digits) function for every element of the sub list, like this:

i = [round(0 - (p * - log2(p)),4) for p in i ]

P.S. I suppose the p in b was p in i

some help here: https://www.w3schools.com/python/ref_func_round.asp

CodePudding user response：

I found a solution for my problem. I realized that I was using the incorrect variables. The code for the expected result is below:

lista=[]

for p in a.values():
    c = [0 - (s * -log2(s)) for s in p]
    lista.append(c)

I have also edited my question to fix the problem in the question as well.