Some Observations on Huffman Trees

Abstract

Information theory and coding theory are very important in communication engineering. Data is transferred across the network from source to destination via some communication channel. Data streams can be very lengthy. Moreover, if we consider all characters from ‘A’ to ‘Z’ in English, the data stream may or may not contain all possible characters in this character set. Hence, we need not use all the bits used to represent the characters from ‘A’ to ‘Z’. Here, data compression comes into play. Coding theory comes up with various techniques required to create codes for each of the characters in the data stream to represent them with smaller number of bits. Huffman coding is one such important algorithmic technique used to create codes for each of the characters. Apart from compression, they are used in other tasks like cryptography, error detection, and correction. In this paper, our concern is data compression mainly. We recommend to club different techniques like use of special characters, contexts in character stream together with Huffman coding to perform effective compression.