On the Entropy of Written Spanish

This paper reports on results on the entropy of the Spanish language. They are based on an analysis of natural language for n-word symbols (n = 1 to 18), trigrams, digrams, and characters. The results obtained in this work are based on the analysis of twelve different literary works in Spanish, as well as a 279917 word news file provided by the Spanish press agency EFE. Entropy values are calculated by a direct method using computer processing and the probability law of large numbers. Three samples of artificial Spanish language produced by a first-order model software source are also analyzed and compared with natural Spanish language. Submitted to the IEEE Transactions on Information Theory