Avoiding NaN Values When Using a Loop

I'm trying to find entropy from a Probability Mass Function (PMF) using the following equation:

Where is a PMF value and m is the amount of values of the PMF.

From there I created this code:

entropy = 0;
pmf = [0.1, 0.015, .. , 0.02, 0.1];
% Summation
for l = 1:length(pmf)
    entropy = entropy + pmf(l) * log2(pmf(l));
    fprintf('Entropy Val %d: %f.20 \n', l, entropy)
end
entropy = entropy * -1;

But the entropy value started to give out a NaN value when . Is there a way to avoid that?

Answer 1

[Matt J]

entropy=-cumsum(pmf.*log2(pmf),'omitnan');

fprintf('Entropy Val %d: %f.20 \n', l:numel(entropy), entropy)

[Jihan Abbas]

Thanks