Clustering the science citation index ® using co-citations

摘要: Abstract  Earlier experiments in the use of co-citations to cluster theScience citation Indey (SCI) database are reviewed. Two proposed improvements methodology introduced: fractional counting and variable level clustering with a maximum size limit. Results an experiment using 1979SCI described comparing new methods those previously employed. It is found that helps reduce bias toward high referencing fields such as biomedicine biochemistry inherent integer count threshold, increases range subject matters covered by clusters. Variable clustering, on other hand, recall measured percentage highly cited items included concluded two used combination will improve our ability generate comprehensive maps science envisioned byDerek Price. This topic be discussed forthcoming paper.