Is There any method can compress data of almost same probability?

the 4bits value of a samll portion like bellow.


4Bits Value   occurrence times
iMatchValue[0] =6
iMatchValue[1] =5
iMatchValue[2] =2
iMatchValue[3] =5
iMatchValue[4] =8
iMatchValue[5] =9
iMatchValue[6] =4
iMatchValue[7] =5
iMatchValue[8] =8
iMatchValue[9] =4
iMatchValue[10]=2
iMatchValue[11]=6
iMatchValue[12]=4
iMatchValue[13]=8
iMatchValue[14]=8
iMatchValue[15]=8

"maplewang" <maplewang@gmail.com> wrote in message
news:1117455830.619850.106000@z14g2000cwz.googlegroups.com...
> the 4bits value of a samll portion like bellow.
>
>
> 4Bits Value   occurrence times
> iMatchValue[0] =6
> iMatchValue[1] =5
> iMatchValue[2] =2
> iMatchValue[3] =5
> iMatchValue[4] =8
> iMatchValue[5] =9
> iMatchValue[6] =4
> iMatchValue[7] =5
> iMatchValue[8] =8
> iMatchValue[9] =4
> iMatchValue[10]=2
> iMatchValue[11]=6
> iMatchValue[12]=4
> iMatchValue[13]=8
> iMatchValue[14]=8
> iMatchValue[15]=8
>

sorted one gets:
9 8 8 8 8 8 6 6 5 5 5 4 4 4 2 2
((((8 6) (5 6)) ((8 8) (8 (4 (2 2))))) ((9 (5 5)) (8 (4 4))))

92*4=368 total

32 24 20 24 32 32 32 20 12 12 27 20 20 32 20 20
379 total

ok, my tree was probably made badly or figured wrong, as I get inflation.

maybe arithmatic coding would do better though, I don't know...

Similar probability distribution is a property of random uncompressible
data. If you data is from a random source, trying to compress it will
be wasting time.

Otherwise, try higher-order models which might give more skewed and
thus more compressible probability distribution.

Sachin Garg [India]
http://www.sachingarg.com

maplewang wrote:
> the 4bits value of a samll portion like bellow.
>
>
> 4Bits Value   occurrence times
> iMatchValue[0] =6
> iMatchValue[1] =5
> iMatchValue[2] =2
> iMatchValue[3] =5
> iMatchValue[4] =8
> iMatchValue[5] =9
> iMatchValue[6] =4
> iMatchValue[7] =5
> iMatchValue[8] =8
> iMatchValue[9] =4
> iMatchValue[10]=2
> iMatchValue[11]=6
> iMatchValue[12]=4
> iMatchValue[13]=8
> iMatchValue[14]=8
> iMatchValue[15]=8