(concluded in 2016)
The LibRank project engaged in the analysis and optimisation of search results in library information systems. It was funded for two years by the German Research Foundation (DFG).
Given the fact that the search habits even of scientists are characterised by the heavy use of commercial search engines, the project aimed to analyse resp. implement the effects and adaptions for library information systems.
One instance of these search habits is the implicit acceptance of the first results for a query as the most important or the most relevant.
LibRank studied these user expectations, taking into account systemic ranking factors such as timeliness, quality, popularity or availability of a search result.
Analysis and modeling of ranking factors
The ranking factors analysed and modelled in the information science-based part of the project were subjected to a tool-supported, systematic and user group-related evaluation. Exemplary test sets for data and search results provided the basis for the assessment of the ranking parameters and algorithms thus developed. These parameters would later enter into the software environment of an actual academic portal – EconBiz, developed and maintained by the ZBW – and become an integral part of the portal based on open source software (VUFind/SOLR).
LibRank addressed all managers and developers of retrieval-based scientific information systems. The findings are available to the interested community of developers for sustainable re-use.
The research project was a joint effort with the Hamburg University of Applied Sciences (HAW). Professor Klaus Tochtermann (ZBW) and Professor Dirk Lewandowski (HAW) were the scientific leaders of the project.