Character Networks and Book Genre Classification

We compare the social character networks of biographical, legendary and fictional texts, in search of statistical marks of historical information. We examine the frequency of character appearance and find a Zipf Law that does not depend on the literary genera and historical content. We also examine global and local complex networks indexes, in particular, correlation plots between the recently introduced Lobby (or Hirsh \(H(1)\)) index and Degree, Betweenness and Closeness centralities. We also found no relevant differences in the books for these network indexes. We discovered, however, that a very simple index based in the Hapax Legomena phenomenon (names cited a single time along the text) that seems to have the potential of separating pure fiction from legendary and biographical texts.