参考链接：
https://segmentfault.com/a/1190000011561822
https://www.cnblogs.com/xubenben/p/3426646.html
http://www.echojb.com/linux-unix/2016/08/17/187428.html
算数编码是熵编码的一种，是基于数据中字符出现的概率，给不同字符以不同的编码。
编码过程：将字符映射到 [0,1) 的区间的一个数。
直接贴代码：

#include <math.h>
#include <string.h>
#include <stdlib.h>

char inStr[100], chSet[20]; //输入字符串和字符集
float P[20];                    //每个字符的概率
float pZone[20];              //概率区间
int strLen;         //输入字符串长度
int chNum;          //字符集中字符个数
int binary[100];
float infoLen;          //信息量大小



void compress();        //编码函数
void uncompress();      //解码函数

int main()
{
    
    int i,j;
            

    printf("input the length of char set:\n");
    scanf("%d", &chNum);
    getchar();

    printf("input the char and its p\n");
    for (i=0; i < chNum; i++) {
        printf("input char: ");
        scanf("%c", &chSet[i]);
            
        getchar();
        //printf("sssss%c  ", chSet[i]); 
        printf("\ninput its p: ");
        scanf("%f",&P[i]);
        getchar();
        printf("\n");
    }
    /***********   test  ************
    for (i = 0; i < chNum; ++i)
         printf("%c<-------------->%f\n", chSet[i], P[i]);
    */
    /************ 计算概率区间 **************/
    pZone[0] = 0;
    for (i=1; i < chNum; ++i)
        pZone[i] = pZone[i-1] + P[i-1];




    printf("input the string \n");
    gets(inStr);
    strLen = strlen(inStr);
    /************* test ***************/
    printf("the string is: \n");
    puts(inStr);

    printf("*********** compress **************\n");
    compress();
    printf("\n*********** uncompress **************\n");
    uncompress();
    return 0; 

}


void compress()
{
    float low = 0, high = 1;
    float L, H, zlen = 1;
    float cp;       //输入字符的概率
    float result;       //结果
    int i, j;
    

    for (i=0; i < strLen; i++) {
        for (j=0; j < chNum; j++) {
            if (inStr[i] == chSet[j]) {
                
                //cp = P[j];
                //L = pZone[j];
                low = low + zlen * pZone[j];
                zlen *= P[j]; 
                break;
            }
        }
        //low = low + zlen * L;
        //zlen *= cp; 
        
    }
    result = low;
    
    printf("the result is %f\n", result);
    
    infoLen = log(1/zlen) / log(2);     //计算香农信息量

    if(infoLen > (int)infoLen)
        infoLen = (int)infoLen + 1;
    else
        infoLen = (int)infoLen;

    /********** 转二进制 *************/

    for (i=0; i < infoLen; i++) {
        result *= 2;
        
        if (result > 1) {
            result = result - 1;
            binary[i] = 1;

        } else if (result < 1) {
                binary[i] = 0;

        } else {
            break;
        }

    }

    if (i >= infoLen) {
        for (j=i; j >= 1; j--) {        
            binary[j-1] = (binary[j-1]+1)%2;
            if (binary[j-1] == 1)
                break;  
        }
    }

    printf("****************** the compress result*****************\n");
    for (j=0; j < i; j++)
        printf("%d ", binary[j]);
}

void uncompress()
{ 
    int i,j;
    float w = 0.5;
    float deResult=0;
    
    float newLow,newLen;
    float low=0,zlen=1;


    /*************** binary to ten *******************/
    for (i=0; i < infoLen; i++) {
        deResult += w*binary[i];
        w *= 0.5;
    }
    printf("uncompress to ten:%f\n", deResult);

    printf("uncompress result:\n");
    for (i=0; i < strLen; i++) {
        for (j=chNum; j > 0; j--) {
            newLow = low;
            newLen = zlen;
                newLow += newLen * pZone[j-1];      
            newLen *= P[j-1];   
            if (deResult >= newLow) {
                low=newLow;
                zlen=newLen;
                printf("%c ",chSet[j-1]);
                break;
            }
        }
    }
    
}