Yurii Orlov
Statistical analysis in literature
The problems of texts classification and identification by authors, genres and other attributes are studied with the use of letter distribution function. The method is based on the kinetic approach to the evolution of empirical non-stationary distribution function. The spectrum of evolution operator is constructed for each text or author. The accuracy of various methods of author identification is discussed. For the problem of text uniformity analysis the method of so-called horizon statistics is used for sequence of distances between the same letters.
| 28.02.2013 | |