Preferential Models of Query by Navigation

This article can be seen as integrating nonmonotonic reasoning and information retrieval. Searching is realized via navigating through an information space called a hyperindex. The user preferences suggested by the path are represented as defaults. and/or preclusion relationships. The semantics of navigation paths are defined in the style of model preference logic. Some information retrieval related properties of these semantics are given. Sound inference rules corresponding to these semantics are also provided. These rules may be used to infer descriptors which are consistent with the preferences for information inherent in the user's navigation path. Such inferences can be used for query expansion or to dynamically alter the information space through which the user is browsing. 1 Introduction The information retrieval (IR) problem can be described as the quest to find the set of relevant information objects corresponding to a given information need, which is represented by a reque..