Abstract

A User Interface Agent (UIA) is a software entity that assists its users in effectively using the user interface software. In electronic commerce applications, the user interface software plays an important role. Users of different kinds and backgrounds will use this software with different expectations. Traditional studies of user interface technology have shown the existence of a "gap" between what the user interface software can actually do and the users' expectations about it. Agent technology, in the form of personalized user interface agents, can help narrow this gap. In this thesis, the supports that software agents can provide in various stages of consumers' shopping process are introduced. Roles of the user interface agent and contents of the user model are analyzed. Also, this thesis explores the use of a user interface agent to provide personalized service to the user by knowing the user's preferences. The UIA maintains a knowledge base about the user, based on a user model. Such user models can be constructed by taking application dependent and application independent attributes into account. The user model is used to deal with incompleteness and ambiguity that might arise in the user's search query, to help the UIA decide when and what kind of proactive information to present to the user, or to enable the UIA to filter the search response set. The UIA assists the user by "reasoning" effectively based on a set of rules that captures the user's needs, preferences and the way he is interacting with the software system. A prototype for UIA has been implemented on the Java platform for electronic shopping. A set of test cases demonstrates that employment of the user model for the UIA improves the effectiveness of the user interface in electronic commerce applications