Huffman and Average Length

问题

I need Matlab code that solves the example problems below. According to the probability values of the symbols I have given, the huffman code will find its equivalent, step by step. If you help me, i will be very happy. I've put examples of this below. All of them have obvious solutions.

For example:

Symbol        Probability
a1                enter probability value input
a2                enter probability value input
a3                enter probability value input
a4                enter probability value input
a5                enter probability value input
a6                enter probability value input
.                      .
.                      .
an                enter probability value input
Output:
a1                1
a2                011
a3                01101
a4                10000
a5                11001
a6                10010
.                    .
.                    .
.                    .
an                 .
L average= .............................

example 1

example 2

example 3

example 4

My tried code block:
function [huffcode,n]=huffmancode(p);
if min(p)<0
    error('Negative element cannot be in a probability vector')
    return
else if abs(sum(p)-1)>1.e-12
        error('Sum of input probability is not 1')
        return
    end
    
    [psort,pord]=sort(p);
    
    n=lenght(p);
    q=p;
    
    for i=1:n-1
        [q,1]=sort(q);
        m(i,:)=[1(1:n-i+1),zeros(1,i-1)];
        q=[q(1)+q(2),q(3:end),1];
    end
    
    Cword=blanks(n^2);
    Cword(n)='0';
    Cword(2+n)='1';
    
    for i1=1:n-2
        Ctemp=Cword;
        idx0=find(m[n-i],:)==)*n;
        Cword(1:n)=[(idx0-n+2:idx0) '0'];
        Cword(n+1:2*n)=[Cword(1:n-1 '1'];
            for i2=2:i1+1
            idx2=find(m(n-i1,:)==i2);
            Cword(i2*n+1:(i2+1)*n)=Ctemp(n*(idx2-1)+1:n*idx2);
            end
    end
    
    for i=1:n
        idx1=find(m(1,:)==i);
        huffcode(i,1:n)=Cword(n*(idx1-1)+1:idx1*n);
    end
  .................................................................................................................................................................................................................................
p=[0.4 0.3 0.1 0.1 0.06 0.04];
[huffcode,n]=huffmancode(p);
entropy=sum(-log(p)*p')/log(2);
display(['symbol','-->',' codeword', Probability'])
for i=1:n
    codeword_length(i)=n-length(find(abs(huffcode(i,:))==32));
    display(['x',num2str(i),' -->',huffcode(i,:),' ',num2str(p(i))]);
end
codeword_Lenght
avg_length=codeword_Length*p';
display(['Entropy = ', num2str(entropy)])
display(['Average codeword length = ', num2str(avg_length)])
....................................................................................................................

for detailed explanation: https://www.programmersought.com/article/34164094982/

来源：https://stackoverflow.com/questions/65760255/huffman-and-average-length

标签

matlab

huffman-code