Title: Automatic extraction of knowledge from student essays

Authors: Maria Vargas-Vera, Emanuela Moreale

Addresses: Knowledge Media Institute (KMI), The Open University, Milton Keynes, MK7 6AA, UK. ' Knowledge Media Institute (KMI), The Open University, Milton Keynes, MK7 6AA, UK

Abstract: This paper presents a characterisation of argumentation in student essays and analyses patterns for extracting knowledge from them. Having analysed their complexity in light of the kinds of logic that may be used in an automatic argumentation extraction system, the main characteristic of these patterns appears to be the polymorphism of the pattern variables. Therefore, systems that learn patterns automatically ought to be able to generate many-sorted logic formulae, so that polymorphic types may be associated to the extraction slots (or – equivalently – to a logic formulae). An analysis of existing (pattern learning) systems was carried out to gauge the possibility of using them within our framework. However, we concluded that none of the existing systems can handle our requirements. Finally, we present our vision of an agent-based student portal as the front-end of a system that can locate argumentation links in a student essay and integrates with related educational services.

Keywords: automatic extraction; knowledge extraction; student essays; agent-based student portals; agent-based systems; multi-agent systems; pattern learning; argumentation links; educational services; information extraction; machine learning; knowledge management.

DOI: 10.1504/IJKL.2005.008354

International Journal of Knowledge and Learning, 2005 Vol.1 No.4, pp.318 - 331

Published online: 07 Dec 2005 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article