Wednesday, March 2, 2016

sequences and series - Does the Shannon Entropy always exist (even for infinite distributions)?

Let p:N[0,1] be a probability distribution over the naturals.




The Shannon Entropy is:



H=n=0p(n)log2p(n)



Does this series always converge?






I tried a little to attack this problem but it's been some time since I evaluate series. My attempt:




We known that n=0p(n)=1, therefore it must be the case that limnp(n)=0, and since limx0xlog2x=0, for any ε>0 there exists only a finite amount of terms greater than ε. The problem is that I can't conclude anything from this, because there are several series in which the terms go to zero but it still diverges.

No comments:

Post a Comment

analysis - Injection, making bijection

I have injection f:AB and I want to get bijection. Can I just resting codomain to f(A)? I know that every function i...