Character Networks and Book Genre Classification

Abstract: We compare the social character networks of biographical, legendary and fictional texts, in search for marks of genre differentiation. We examine the degree distribution of character appearance and find a power law that does not depend on the literary genre or historical content. We also analyze local and global complex networks measures, in particular, correlation plots between the recently introduced Lobby (or Hirsh $H(1)$) index and Degree, Betweenness and Closeness centralities. Assortativity plots, which previous literature claims to separate fictional from real social networks, were also studied. We've found no relevant differences in the books for these network measures and we give a plausible explanation why the previous assortativity result is not correct.