Guy Saward - Research

Projects

• OPAC - Overcoming Problems of Academic Content

  The OPAC project aims to overcome the current situation where investments in high-quality academic information sources and content are ignored by students and other users who favour lower quality, but easily accessible information retrieved via Internet search engines.

  To achieve this culture change, the project aims to overcome previously researched problems with: speed and ease of searching; understanding of information search and retrieval concepts; and the selection and use of appropriate sources.

• FIN - Feasibility of Integrating Naviguidance

  The aim of the Fully Integrated Naviguidance (FIN) project has been to investigate how web-site navigation can be improved through the addition of "naviguidance" - a way to guide users as they navigate through large or complex collections of information.

  This has been achieved through the creation of a number of software systems in conjunction with British Airways, Boots and Tesco. These case studies have enabled us to meet our objectives in assessing technical, user and business benefits.

Papers

• Changing Designs for Internet Information System Navigation, Usability and Acceptance
  submitted for publication
  
  ABSTRACT: System designs constantly change as user requirements and technology evolve. Designers need tools that ensure design changes deliver long-term user benefit rather than cosmetic, quick fixes. We present a case study on the evaluation of a number of design changes to the intranet of a global airline. Both are motivated by information foraging theory and the Technology Acceptance Model. The design changes include a new form of guided navigation designed to increase information scent. The evaluation incorporates new metrics used measure resulting changes in user intention and task/strategy fit that result from changes in scent. The counter-intuitive results provide insight into some of the drawback inherent in small scale usability studies.
  
  Full paper - PDF version
   
• Assessing Usability through Perceptions of Information Scent
  The 10th International Software Metrics Symposium, IEEE Press
  
  ABSTRACT: Information scent is an establish concept for assessing how users interact with information retrieval systems. This paper proposes two ways of measuring user perceptions of information scent in order to assess the product quality of Web or Internet information retrieval systems. An empirical study is presented which validates these measures through an evaluation based on a live e-commerce application. This study shows a strong correlation between the measures of perceived scent and system usability. Finally the wider applicability of these methods is discussed.
  
  Full paper - PDF version
   
• A Case Study of Confidence in Internet Information Retrieval
  26th European Conference on Information Retrieval, Sunderland, UK
  
  ABSTRACT: This poster describes a study into the relationship between user knowledge and confidence in Internet information retrieval and its affect on user behaviour and success. The results show a significant link between user expectations and how easy users perceive a task to be. Detailed analysis identifies differences between how confidence is acquired when browsing as compared to searching.
  
  Full paper - PDF version
   
• Macro and Micro Applications of Case-Based Reasoning to Feature-Based Product Selection
  ES2000 Conference on Expert Systems, Cambridge, Springer Lecture Notes in AI
  
  ABSTRACT: This paper examines alternative applications of case-based reasoning to product selection. It is motivated by shortfalls in e-commerce solutions but addresses wider issues in the selection of case representations and reasoning strategies where solution requirements are negotiable and the number of potential features are large. Two approaches are examined: the micro approach uses the traditional CBR nearest-neighbour algorithm to augment product searches, while the macro approach is based on the traditional CBR-cycle but substitutes crisp constraint relaxation for similarity based retrieval. While the discussion is set against the background of a specific case study, the results have a wider applicability across a number of problem domains.
  Full paper - PDF version
   
• A FAB Approach to E-Commerce Knowledge Accessibility Requirements
  
  ABSTRACT: E-commerce sites currently provide poor access to information and knowledge. They neither sell products nor sell the concept of e-commerce. A review of the top 15 e-commerce sites, shows that their success is based on knowledgable customers rather than knowledge delivered through the web. This conclusion is based on the development of knowledge-level accessibility requirements and a features/advantages/benefits sales model. The paper concludes with a discussion of how e-commerce sectors outside books, PC suppliers and travel companies might succeed in selling to the majority of customers rather than relying on knowledgable customers to buy products at discounted prices.
  
  Full paper - PDF version
   
• Using Cases to Model Processes: An Example from the Water Supply Industry
  12th International Conference on Industrial and Engineering Application of AI and Expert Systems, Cairo, Springer Lecture Notes in AI
  
  ABSTRACT: This paper describes the use of cases as a process representation technique and shows how case base reasoning can be used to navigate through, or execute a complex control process. The work is based on a prototype decision support application for a UK water utility company in which two processes were modelled - one diagnostic and one operational. A brief description is given of how processes are represented as cases and how those cases are are used to animate the business process.
  
  Full paper - PDF version
   
• Representing Process Knowledge Using Case-Based Reasoning
  4th UK Workshop on Case-Based Reasoning, Salford
  
  ABSTRACT: Capturing knowledge about reasoning processes is a principle theme of AI. The techniques for doing this range from rule-based systems that engineer the process in great detail, to case-based retrieval that uses simple, data driven algorithms. The Case-Based Process approach described in this paper synthesises these two extremes by describing and demonstrating how cases can be used to represent process knowledge. Chains of inferences are captured as discrete cases resulting in a coherent knowledge base with a high level of clarity and explanatory power. We describe how process descriptions can be translated into cases and how that knowledge is animated by an incremental case retrieval process. The degree of engineering associated with CBP is compared to the traditional CBR and RBS approaches, with particular regard to the knowledge engineering bottleneck. The capabilities required to implement scaleable CBP applications are discussed, based on a number of industrial applications. These applications demonstrate the importance and benefits for capturing process knowledge for use in decision support tools in a help desk environment and for Internet based self-service assistants.
  
  Full paper - PDF version
   
• Navigating Heterogeneous Knowledge Sources in a Customer Service Environment
  2nd International Workshop on Innovative Internet Information Systems, Copenhagen
  
  ABSTRACT: Web-based searching utilises both static and dynamic knowledge structures but does not link the two. We present an approach based on navigating heterogeneous knowledge in a customer service environment that combines dynamic queries with a static subject structure. By utilitising a hierarchy that is tailored to a specific organisation and takes account of its knowledge management goals we can produce a more effective search.
  
  Full paper - PDF version
   
• Modelling Intelligent Agents and Assistants in a Call Centre Environment
  7th European Congress on the Use of Intelligent Techniques, Aachen
  
  ABSTRACT: The call centre environment is a fertile area for the application of intelligent software agents. This paper examines a number of problems that can be addressed through the application of agents built using a knowledge-based approach. Case-based reasoning is used as the basic knowledge representation within a 3-tier architecture. The use of cases as a knowledge representation technique for agents within a call centre environment over comes some of the issues typically associated with knowledge-based agents.
  
  Full paper - PDF version
   

Last updated on 01/10/2004
By Guy Saward
Email: g.r.saward@herts.ac.uk