On the Connection Between Citation-based and Topical Relevance Ranking: Results of a Pretest using iSearch

In Information Retrieval a lot of work and effort was put into the exploitation of textual information. Alternative approaches like citation search strategies make use of structural information that can be found in digital libraries or scientific information systems. One dataset containing structural information and everything needed for an IR evaluation is the iSearch test collection. Given this test collection we investigate its suitability with a special interest in bibliometric methods like citation analysis. By using co-citation recommendations we look for a topical connection between the seed document (which is known to be relevant to the given topic) and the recommended documents.