No zero match browsing of hierarchically categorized information entities

Abstract

The two main ways to search for information in electronic document collections are by using free-text retrieval search engines or browsing information that has been organized into predefined organizational structures. However, each of these approaches has limitations. Using word or phrase search, users are faced with a compromise between overly broad searches returning an excessive amount of information or overly narrow searches that may fail to return relevant information. Browsing organizational structures is dependent on the user's knowledge of the structures, and a user may find it difficult to refine searches. This paper introduces a user interface based approach to the browsing of hierarchically organized information entities that avoids these problems by allowing the incremental narrowing down of a set of search results and by pruning the organizational structure after each user selection to show the consequences of the selection. The effect is to present to the user at all times only that part of the organizational structure that will lead to a nonnull selection. The approach is called no zero match (NZM) browsing. The paper presents the computational basis of NZM browsing before describing a trial implementation of the approach and presenting three case studies, which represent common search situations in an engineering context.