Science Models as Value-Added Services for Scholarly Information Systems

TL;DR

This paper explores how science models can serve as value-added services in scholarly information systems, enhancing retrieval quality by representing scientific structures and activities, and validating their effectiveness through empirical evaluation.

Contribution

It introduces science models as innovative search services in scholarly IR, demonstrating their positive impact on retrieval performance and understanding of scientific structures.

Findings

Science models improve retrieval quality

Models effectively represent scientific phenomena

IR perspective enhances understanding of scholarly activities

Abstract

The paper introduces scholarly Information Retrieval (IR) as a further dimension that should be considered in the science modeling debate. The IR use case is seen as a validation model of the adequacy of science models in representing and predicting structure and dynamics in science. Particular conceptualizations of scholarly activity and structures in science are used as value-added search services to improve retrieval quality: a co-word model depicting the cognitive structure of a field (used for query expansion), the Bradford law of information concentration, and a model of co-authorship networks (both used for re-ranking search results). An evaluation of the retrieval quality when science model driven services are used turned out that the models proposed actually provide beneficial effects to retrieval quality. From an IR perspective, the models studied are therefore verified as expressive conceptualizations of central phenomena in science. Thus, it could be shown that the IR perspective can significantly contribute to a better understanding of scholarly structures and activities.