Extraction of Message Sequence Charts from Narrative History Text

摘要: In this paper, we advocate the use of Message Sequence Chart (MSC) as a knowledge representation to capture and visualize multi-actor interactions their temporal ordering. We propose algorithms automatically extract an MSC from history narrative. For given narrative, first identify verbs which indicate then dependency parsing Semantic Role Labelling based approaches senders (initiating actors) receivers (other actors involved) for these interaction verbs. As final step in extraction, employ state-of-the art algorithm temporally re-order interactions. Our evaluation on multiple publicly available narratives shows improvements over four baselines.