Evaluating Relevance Feedback and Display Strategies for Searching on Small Displays

Abstract

Searching information resources using mobile devices is affected by displays on which only a small fraction of the set of ranked documents can be displayed. In this study we explore the effectiveness of relevance feedback methods in assisting the user to access a predefined target document through searching on a small display device. We propose an innovative approach to study this problem. For small display size and, thus, limited decision choices for relevance feedback, we generate and study the complete space of user interactions and system responses. This is done by building a tree – the documents displayed at any level depend on the choice of relevant document made at the earlier level. Construction of the tree of all possible user interactions permits an evaluation of relevance feedback algorithms with reduced reliance on user studies. From the point of view of real applications, the first few iterations are most important – we therefore limit ourselves to a maximum depth of six in the tree.