Information Extraction for Modelling Screenplay Evolution of Star Wars Fiction

Аннотация

This work is an attempt towards automated fictional network analysis using character-level co-occurrence networks and their embeddings to study clustering screenplays in inconsistent and very ambiguous series of Star Wars fiction. We use data from 353 Star Wars books, separated by screenplay dates and series types (novels, short stories, Legends and Canon) to provide new insights in the topic of screenplay evolution and similarity using network-based approaches.