Development and research of adaptive data compression methods based on linear fibonacci form

Abstract

A fundamentally new data compression approach, which is based on the optimizing properties of Fibonacci numbers lies in the fact that the figures are considered as positive whole numbers and presented by a linear Fibonacci form, was investigated. Formation features of the numerical data source model were examined. The effect of the length of data blocks of the compressed file on the compression ratio was studied. Changing the number of bytes in the block provides the formation of different data source models. Ability to change the data source model allows to choose a model that provides the greatest compression ratio for a given encoding rule. Analysis of the results has shown that the effect of the data block length on the compression ratio is different for different file types. For some types, the greatest compression ratio is achieved when the block length is 100 bytes, and the ratio decreases with the increased length. For other file types, the effect of the data block length has a wave pattern (the ratio repeatedly increases and decreases with the increased length), and for certain types of files, the dependence of the transformed data on the data source model used is negligible. Low compression ratios and no compression for certain file types are caused by the fact that the data modeling used does not ensure the formation of numbers that are compactly presented by the linear Fibonacci form. To eliminate this shortcoming, two adaptive data compression methods, based on the linear Fibonacci form, which involve using a set of numerical data source models were proposed and investigated. These models are based on the maximum value of the numerical equivalents of the ASCII codes of bytes that make up the block. Adaptation enhances the compression ratio compared to non-adaptive compression method based on the linear Fibonacci form.