Syllabic quantity patterns as rhythmic features for Latin authorship attribution

It is well known that, within the Latin production of written text, peculiar metric schemes were followed not only in poetic compositions, but also many prose works. Such patterns based on so-called syllabic quantity, i.e., length involved syllables, and there substantial evidence suggesting that certain authors had a preference for over others. In this research we investigate possibility to employ quantity as base deriving rhythmic features task computational authorship attribution texts. We test impact these when combined with other topic-agnostic features. Our experiments, carried out three different datasets, using two machine learning methods, show are beneficial discriminating among authors.