Locating information that one needs in the ever-increasing amount of electronic data requires sophisticated searching tools. Unfortunately, the sophistication of information retrieval systems has not kept up with the quantity of information now available. This research addresses this issue by synthesizing the research disciplines of user modeling, information retrieval, software agents, and adaptive interfaces into a theoretical framework to improve existing information retrieval systems. An application, known as the Agent Improved Information Retrieval System (AI2RS) was developed based on this model. To properly test the AI2RS system, a new evaluation metric, sufficiency, was developed that is more robust and user-focused than traditional information retrieval evaluation metrics. Ergonomic testing of the agent was also conducted. Implications of this type of application are explored and future research directions are suggested. |