Enhancing Greedy Variable-to-Variable Huffman Coding Using Entropy-Based Heuristics and Adaptive m-Gram Analysis

Abstract

The paper focuses on enhancing the Greedy Variable-to-Variable Huffman Cod- ing algorithm by addressing its limitations through the use of a heuristic approach. The current greedy algorithm has several shortcomings, such as lack of context- awareness, no consideration of data structure or entropy, and no use of historical m-gram data. These issues limit its ability to make optimal compression decisions. The paper aims to improve compression efficiency by incorporating local en- tropy estimates, leveraging historical m-gram usage, and developing an adaptive decision-making framework. Our approach improves sequence selection by dynam- ically balancing symbol predictability, codeword efficiency, and redundancy reduc-
tion.