APPROACHES FOR THE DIGITAL PROFILING OF ACTIVITIES AND THEIR APPLICATIONS IN DESIGN INFORMATION PUSH
Author: Campbell, Duncan Russell
Supervisor: Culley, Steve; McMahon, Chris
Institution: University of Bath, Department of Mechanical Engineering
This thesis describes research carried out to investigate and develop ‘information push’ technologies. Essentially, information push approaches introduce autonomy in the dissemination of electronically stored information between computer users. The work has been oriented to establish benefits in the general area of information management, but with an emphasis on issues faced in the field of engineering design. Airbus UK, a stakeholder in the research, provided the impetus for the research topic having highlighted current information management problems in their organisation. In the literature review, an endeavour is made to provide a unified overview of theory and technologies geared towards information push. This includes the consolidation and critique of literature in the fields of Information Retrieval, Artificial Intelligence, Human Computer Interaction and other domains. The results reveal that there are many avenues and approaches for implementing push-based systems. As a result, a large number of combinatory approaches are possible, but many have yet to be fully reviewed or researched. This finding is further substantiated from a state-of-the-art review conducted on existing academic and commercial systems. The research has resulted in a number of contributions and developments. A key contribution is the development of a model identifying the primary modes of information push which help to define the scope of the research field. A further contribution is made with the generation of a range of approaches specifically developed with applications in design information management in mind—highlighting numerous avenues worthy of further research. The most promising approach arising from the early research was extended, and a supporting software prototype was developed. The approach, which supports a particular type of information push, uses captured data to identify relationships between documents used by an individual or group, representing their value in a given context – that may relate to a specific information need or activity. A final contribution is made through the testing and evaluation of this approach in industrial and academic settings. The main results indicate that it provides a method for the establishment of a novel type of document similarity matrix which can be usefully exploited in the development of push technologies. It is clear from the research that the paradigm of information push has much potential to be of benefit in certain working domains and circumstances. However, it is also clear that associated supporting technologies are in their infancy. It is concluded that further advancement of push technology is reliant on more detailed and sophisticated capture and interpretation of computer user’s interactions – specifically in relation to interactions with electronic information. This key point is identified as the main barrier in the development of more powerful systems delivering relevant information to computer users, improving efficiency in information intensive working environments.
Keywords: Information management, Information set, Information stream, Information Retrieval, Information Filtering, Collaborative Filtering, Content filtering, Economic filtering, Agent, Multi-Agent-Systems, RSS