الفهرس 
AcknowledgementOnly 14 pages are availabe for public view
 |  | from | 73 |  |  |
| | | from | 73 | | |
Abstract
The classic proposal believes that the new numerical representation of data does not affect entropy, as data remains the same. Thus, hypothesis of this thesis can be exposed as an investigation and performance analysis of the effect of representing the dataset using “base64” on arithmetic compression ratio where arithmetic is an entropy compression technique.
A Compression tool is developed using C# programming language. It has been used to perform encoding and decoding functions using base64 technique, perform compression and decompression using arithmetic compression technique and calculate the entropy of received and generated files.
In order to obtain valid and reliable results, the proposed solution uses a benchmark dataset which is supported by many universities.
The results insure that the compression ratio is affected positively and negatively according to the entropy of the data subset. “base64” technique is used before applying the arithmetic technique, where “base64’” is just a numerical representation. It results in a change of entropy scheme of file regarding the increasing of the symbols probability from 1/256 in ASCII mode in original file into 1/64 in “base64” representation. In turn, this affects on repetition of symbols in message which affects directly to entropy.