How to optimize Shannon entropy calculation function?

Question

I did this function to calculate the entropy of Shannon in a string array, this vector is filled with Ips addresses, I would like to optimize this function since for cases where the input vector is too large, the calculation time is too high.

function [entropy] = ShannonEntropy(IPs)
%% ENTROPY
 % Function calculates Shannon entropy for a string vector

 %% Copy vector IPs
 searchIP = IPs; 

%% Determines the size of the vector
size = length(IPs);

%% Initialize vector 
 % P will be used to calculate probability
P = zeros(size, 1);

%% Finds occurrences
for i=1:size
 for j=1:size
  if (searchIP(i) == IPs(j) && IPs(j) ~= "-1") 
   P(i) = P(i)+1;
   IPs(j) = "-1";
  end
 end
end

%% Calculate probabilities
P = P()/size;

%% Replace 0 with 1
 % ie: log 0 = NaN, log 1 = 0
P(P == 0) = 1;

%% Initialize vector
partialEntropy = zeros(size, 1);

%% Calculates entropy
for i=1:size
 partialEntropy(i) = (P(i)).*log(P(i));
end

%% Entropy sum
entropy = -sum(partialEntropy);