Spoken language classification using hybrid classifier combination

Article type: Research Article

Authors: Garfield, Sheila^{; *} | Wermter, Stefan | Devlin, Siobhan

Affiliations: Centre for Hybrid Intelligent Systems, School of Computing and Technology, University of Sunderland, St. Peter's Way, Sunderland SR6 0DD, UK

Correspondence: [*] Corresponding author. sheila.garfield@sunderland.ac.uk

Abstract: In this paper we describe an approach for spoken language analysis for helpdesk call routing using a combination of simple recurrent networks and support vector machines. In particular we examine this approach for its potential in a difficult spoken language classification task based on recorded operator assistance telephone utterances. We explore simple recurrent networks and support vector machines using a large, unique telecommunication corpus of spontaneous spoken language. The main contribution of the paper is a combination of techniques in the domain of call routing. First, we find that simple recurrent networks perform better than support vector machines for this task. Second, we claim that the combination of simple recurrent networks and support vector machines provides slightly improved performance compared to the performance of either simple recurrent networks or support vector machines.

Keywords: classification, spontaneous language, dialogue, recurrent neural networks, support vector machines

DOI: 10.3233/HIS-2005-2102

Journal: International Journal of Hybrid Intelligent Systems, vol. 2, no. 1, pp. 13-33, 2005