Empirical Text Mining for Genre Detection

In this paper, we report on a preliminary study we carried out for identifying patterns that characterize the genre type of Greek texts. In the course of our study, we address four distinct genre types, we record their observable stylistic elements and we indicate their exploitation for automatic genre-based document classification. The findings of our study demonstrate that texts contain lexical features with discriminative power as far as genre is concerned, however modeling those features so that they can be explored by computer-based applications is still in early stages.