MATLAB实现算术编码

str1 = ['d', 'a', 'c', 'a', 'b'];
%统计字符串中的字符种类，调用函数，放入数组b中
b = unique(str1); 
%统计每种字符的个数，放入数组a中
for i = 1 : length(b)
    a(i) = length(strfind(str1, b(i)));
end
%排序，若数组b中的第一个元素为字符‘a’, 则数组a中的第一个元素为字符‘a’出现的次数
for i = 1 : length(a)-1
    for j = i+1 : length(a)
        if b(i) > b(j)
            t1 = a(i); a(i) = a(j); a(j) = t1;
            t2 = b(i); b(i) = b(j); b(j) = t2;
        end
    end
end
%统计每种字符出现的频率，放入数组fre中
for i = 1 : length(a)
    fre(i) = a(i) / sum(a);
end
%根据频率分配概率范围，范围高值放入H中，范围低值放入L中
Li(1) = 0; Hi(1) = fre(1);
for i = 2 : length(fre)
    Li(i) = Hi(i-1);
    Hi(i) = Li(i) + fre(i);
end
%编码
L = 0; H = 1; %初始化
for i = 1 : length(str1)
    W = H - L; 
    d = strfind(b, str1(i)); %找到字符串中的字符，在数组d中的角标
    H = L + W * Hi(d);
    L = L + W * Li(d);
end

code = [];
while true
    tmp1 = 0; tmp2 = 0;
    L = 2 * L; H = 2 * H;
    if L >= 1
        tmp1 = 1; L = L - 1;
    end
    if H >= 1
        tmp2 = 1; H = H - 1;
    end
    if tmp1 ~= tmp2
        code = [code '1']; break;
    else
        code = [code tmp1+'0'];
    end
end
code
%译码
N = 0; 
%二进制转化为十进制小数
decode = [ ];
for i = 1 : length(code)
    N = N + (code(i) - '0') * power(2, -1*i);
end
for i = 1 : length(str1)
    for j = 1 : length(fre)
        if N >= Li(j) && N < Hi(j) %查找N属于哪个间隔
            decode = [decode b(j)]; 
            W = Hi(j) - Li(j);  
            N = (N - Li(j)) / W;  %更新N
            break;
        end
    end
end
decode