AAAI Publications, The Twenty-Sixth International FLAIRS Conference

Font Size: 
Automatic Detection of Nominal Entities in Speech for Enriched Content Search
Ricardo A. Calix, Leili Javadpout, Mehdi Khazaeli, Gerald M. Knapp

Last modified: 2013-08-28


In this work, a methodology is developed to detect sentient actors in spoken stories. Meta-tags are then saved to XML files associated with the audio files. A recursive approach is used to find actor candidates and features which are then classified using machine learning approaches. Results of the study indicate that the methodology performed well on a narrative based corpus of children’s stories. Using Support Vector Machines for classification, an F-measure accuracy score of 86% was achieved for both named and unnamed entities. Additionally, feature analysis indicated that speech features were very useful when detecting unnamed actors.

Full Text: PDF