Learning and the structure of citation networks

摘要: The distribution of citations received by scientific publications can be approximated a power law, finding that has been explained “cumulative advantage”. This paper argues socially embedded learning is plausible mechanism behind this cumulative advantage. A model assuming scientists face time trade-off between and writing papers, they learn the papers known their peers, cite know, generates law popularity, shifted for received. two distributions flatten if there relatively more learning. predicted exponent independent average in-(or out-) degree, contrary to an untested prediction reference (Price, 1976). Using publicly available citation networks, estimate share devoted (against producing) given around thirds.