Arithmetic Assistance

Giganews Newsgroups
Subject: Arithmetic Assistance
Posted by:  Einstein (michael…@gmail.com)
Date: Wed, 14 May 2008

I am trying to learn more of the Arithmetic method of compression. So
I laid this model to myself....

I have a 2 in 14 chance to have a compressible result, and a 12 in 14
chance of having a non-compressible result. Statistically speaking of
course. Now I figure the cost to determine will be higher than the
compression saved, I am now convinced of the 'no random recursive
compression' plan... but I am trying to create a basis to do some
compression models, but I am having a hard time with arithmetic
compression.

Would I use a similar to huffman method to determine if I have
compression?

When I do that I have two odds, 1 or 0. There are 4 outcomes. I take
the percentages  of 1+1 * 3, 1+0 * 3, 01 *2 and 00 *1 and divide all
four results to get the compression (Or size increase) of the simple
huffman. If this is the way this math is calculated then that is easy
enough. But is there alternatives? I feel like I am missing something
here...

Replies