Sistema adaptativo para compressão de dados

Abstract

An interesting property is proven for Huffman's tree. Any two elements with weights ai and bi sons of a same father have the following property: if bi ≥ ai then Wm ≤ ai or Wm ≥ bi where Wm is the weight of any node of the tree. Based on this property, an algorithm is developed to dynamically update Huffman's tree as weights of terminal nodes change. Using this algorithm, a model of an adaptive system for data compression is developed. Simulation using many types of data led to interesting results.