Content Coverage of PNAS in 1995 and 2001
Ketan K. Mane, Katy Börner · 2001
This paper reports results of determining changes in content coverage of the Proceedings of the National Academy of Science (PNAS). Two time slices namely 1995 and 2001 were selected as sample data sets. Latent Semantic Analysis [5] was applied to determine semantically similar documents. In addition, co-word analysis was used to examine the occurrence of terms used in titles and keywords of articles. The results were visualized using graph layout algorithms available in the graphical visualization software Pajek [1]. Data Set The utilized data set comprises the complete set of papers in the Proceedings of National Academy of Science for the year 1995 and 2001. For each document the author and publisher assigned keywords and controlled vocabulary Mesh terms from Medline were determined. Statistics on the two datasets are given in Table 1. Obviously papers published in 2001 had little time to receive a significant number of citations. Table 1: Statistics of the PNAS data sets