Ranking research projects by their similarity – a semantic servicein the Hungarian Current Research Information System (HunCRIS)

Abstract

HunCRIS has developed a new semantic application whichallows calculation of the numeric value of similarity between any two projects. The paperdescribes the relevant algorithm by figures and explanations, and compares it to Booleansearch solutions. In Boolean search the key expressions describing projects are searchedamong the items included in the search expression and their sub-terms in the thesaurus. Thehits are typically ordered by non-search-relevant properties. The extent of similarity iscalculated as the correlation between sets of key expressions and their super-terms in thethesaurus that describe the selected projects in the knowledge space. The projects can beranked according to the metrics of similarity, or according to their relevance to a searchexpression. The service is available on the public site of HunCRIS.