Simulating Interactive Information Retrieval: SimIIR: A Framework for the Simulation of Interaction

摘要

Simulation provides a powerful and cost-eu000bective approachrnto explore and evaluate how interactions between a searcherrnand system inruence search behaviour and performance.rnWith a growing interest in simulation and an increasingrnnumber of papers using such an approach, there is a needrnfor a rexible framework for simulation. Thus, we presentrnSimIIR, an open-source toolkit for building and conductingrnInteractive Information Retrieval (IIR) experiments. Thernframework consists of a number of high level components,rnincluding the simulation, the searcher and the system, all ofrnwhich must be confgured. The SimIIR framework providesrna series of interchangeable components. Examples of theserncomponents include the querying strategies (how simulatedrnqueries are formulated) and stopping strategies (the depthrnto which a searcher will examine snippets and documents)rnthat a simulated searcher will employ. We have implementedrnvarious existing strategies so that they can be used by otherrnresearchers to not only replicate and reproduce past experiments,rnbut also create new experiments. This paper describesrnthe SimIIR framework and the diu000berent componentsrnthat can be confgured and extended as required.