In this paper we investigate a general purpose interacive information organization system. teh system organizes documents by placing them into 1-, 2-, or 3-dimensional space based on their similarity and a spring-embedding algorithm. We begin by designing a method for estimating the quality of the organization when it is applied to a set of documents returned in response to a query. We show how the relevant documents tend to clump with each other in space. We proceed by presenting a method for measuring the amount of structure in the organization and we explain how this knowledge can be used to refine the system. we also show that increasing the dimensionality of the organization generally improves its quality. We introduce two methods for modifying the organization based on the information obtained from the user and show how such feedback improves the organization. All the analysis is done off-lien without direct user intervention.
展开▼
