Authors: Yandong Fan
Addresses: Caulfield School of Information Technology, Monash University, 900 Dandenong Rd., Caulfield East, Victoria 3145, Australia
Abstract: Designing a dialogue strategy for speech-enabled mobile commerce is a significant challenge due to the context. This paper introduces a hybrid dialogue strategy to overcome the inflexibility of application-directed interactions while avoiding the significant recognition difficulty of a full user-initiative style. The system uses N-Gram grammars to govern the recognition at the request segment of a dialogue, and employs an application-directed strategy at the clarification discourse segment. The paper also details generating a corpus for the N-Gram grammar through a Case-based Reasoning approach, and constructing application-directed dialogues with decision trees. Our preliminary testing indicates that the strategy is a feasible and effective solution for voice-enabling mobile commerce applications.
Keywords: mobile commerce; speech interface; dialogue strategy; decision trees; N-gram grammar; CBR; case-based reasoning; electronic business; e-business; speech-enabled m-commerce.
International Journal of Electronic Business, 2009 Vol.7 No.2, pp.170 - 189
Published online: 12 Apr 2009 *Full-text access for editors Access for subscribers Purchase this article Comment on this article