Who is the author of the book?

2002/02/01 Elhuyar Zientzia Iturria: Elhuyar aldizkaria

• Softwarea

A new computer study can help identify the author of a book. A computer program recently developed in Italy allows to detect differences between authors that so far could not be seen. This algorithm builds a linguistic tree between 50 different languages. This linguistic differentiation is not difficult in a text, but the differentiation of style is the one that has the most work.

Researchers at the University of La Sapienza analyze the written language as a chain of symbols. In these chains of symbols the entropy is observed, that is, the sequence of repetition of certain symbols. In this way it is possible to summarize the phrases without losing meaning. This is the basis of computer programs that perform data compression. Using this base, researchers can calculate the entropy of two string letters. This distance between entropías is usually greater between texts of different authors than between texts of the same author.