A Dictionary for Minimum Redundancy Encoding

Abstract

Significant advances in iaformation transmission systems have resulted from analysis of the properties of communication channels and transmitted signals. Statistical properties of the information source, however, have usually been ignored, especially in transmission of natural languages. By taking into consideration the statistics of an information source an efficient information transmission system employing minimum redundancy word codes can be developed. The integration of computers as active dements in the system affords opportunities for fuller utilization of the insights of information theory and symbol manipulation techniques. An operating system based upon the above concepts has been developed, programmed for a computer and tested [9]. This paper deals with the problems associated with the information source. The size of a dictionary is considered in Section 2 together with split and integrated dictionary formats. An analysis of an experimental dictionary is presented in Section 3. Section 4 describes the theoretical basis and computer programs for word decomposition and synthesis. Dictionary growth and adaptation are considered in Section 5.