Correlation analysis of the bigrams of the copies of Russian Gospels of the 11th - 14th Centuries

Abstract

With the aid of the correlation analysis the paper considers the degree of closeness of the copies of Russian Gospels of the 11th - 14th century of different types (complete aprakoses, short aprakoses, Four Gospels) to one another, and also the degree of closeness of the Panteleymon Gospel of the 12th century (complete aprakos) to each of the types. The lists of bigrams with the greatest value of T-score whose components are lemmas has been analyzed; the amount of retrievals from each subcorpus has been 300 elements. To reveal the closeness of the lists, the nonparametric statistic r-Spearman was applied to the ranks of bigrams and the parametric statistic r-Pearson to their T-score values. The obtained results indicate the presence of the correlation relationship between the compared arrays of bigrams which has a high statistic probability and also a sufficiently significant degree of their correlation which is characterized either like moderate or noticeable. The assessment of the relationship force between the subcorpora suggests differences in the degree of closeness of the compared arrays of bigrams. In accordance with the rank correlation r-Spearman, the highest closeness is shown by the subcorpus of complete aprakoses and the subcorpus of short aprakoses, the Panteleymon Gospel and complete aprakoses, the lowest by short aprakoses, Four Gospels, the Panteleymon Gospel and Four Gospels (or short aprakoses). In accordance with the correlation r-Pearson the highest closeness is revealed between the complete aprakoses and Four Gospels, the lowest between the complete and short aprakoses. The relationships of the Panteleymon Gospel with Four Gospels and short aparakoses are similar to the relationships of the complete aprakoses with them.