Detecting Events in Narrative Essays

Lam, H.W. and Dillon, T. and Chang, E. and Wongthongtham, P. and Lui, W. 2012. Detecting Events in Narrative Essays, in 26th International Conference on Advanced Information Networking and Application Workshops, Mar 26-29 2012, pp. 363-368. Fukuoka: IEEE.

Source Title

Advanced Information Networking and Application Workshops (WAINA) 2012

Source Conference

2012 26th International Conference on Advanced Information Networking and Application Workshops

DOI

10.1109/WAINA.2012.102

ISBN

978-1-4673-0867-0

School

School of Information Systems

URI

http://hdl.handle.net/20.500.11937/15732

Collection

Curtin Research Publications

Abstract

A narrative can be described as a series of events connected to one another, which are caused by or experienced by the characters of the story. By this definition, an Event can be interpreted as an important part of the story since it is mostly these kinds of statements that move the plot along. The concept proposed in this paper features an Event Detection framework that would eventually allow a grading system to structure the content of an essay into an event flow. The tests show some promising results, with high Precision and Recall values, and an F-measure of over 0.8 in all tests. In addition, the classifier achieves a Matthews Correlation Coefficient average of 0.52. Future work includes integrating Causal Links into the system to detect the relations between Events.