Анастасия Кравцова » Publication

Abstract This paper discusses statistical methods of principal component analysis and discriminant analysis which allow one to reduce a fea- ture space, and considers their application to the text style compar- ison. The observed frequencies of text style features are transformed to their ranks and then by applying nonlinear transformation to nor- mally distributed random variables. The application to the analysis of frequencies of 55 syntactic words in texts taken from the Russian ction of the XIX century is provided.