Use this URL to cite or link to this record in EThOS: http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.547403
Title: ArabChat : an Arabic conversational agent
Author: Hijjawi, Moh'd Hatim Husni
Awarding Body: The Manchester Metropolitan University
Current Institution: Manchester Metropolitan University
Date of Award: 2011
Availability of Full Text:
Access through EThOS:
Abstract:
This thesis details the development of a novel and practical Conversational Agent for the Arabic language called ArabChat. A conversational Agent is a computer program that attempts to simulate conversations between machine and human. In this thesis, the term 'conversation' or 'utterance' refers to real-time chat exchange between machine and human. The proposed framework for developing the Arabic Conversational Agent (ArabChat) is based on Pattern Matching approach to handle users' conversations. The Pattern Matching approach is based on the matching process between a user's utterance and pre-scripted patterns that represents different topics organised through novel scripting structure. ArabChat classifies users' utterances as either question or nonquestion utterances in order to response to an utterance depending on its type (question or non-question). In addition, ArabChat has the ability to reply to an utterance targets many topics at the same time. Moreover, this thesis proposes to use the stemming technique (a process to return a word to its original root) as a pre-processing stage in ArabChat in order to convert the processed utterance's words to stemmed words and then match them with stemmed pre-scripted patterns. This proposal might decrease the number of needed patterns to script a domain to the minimum as discussed in this thesis. Furthermore, two new Arabic stemmers have been proposed, developed and discussed in this thesis In order to assess ArabChat, different techniques have been developed and used to validate its performance. Three experiments with online users have been carried out. The results have shown that ArabChat is effective.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID: uk.bl.ethos.547403  DOI: Not available
Share: