Predictive ranking of computer scientists using CiteSeer data

The increasing availability of digital libraries with cross‐citation data on the Internet enables new studies in bibliometrics. The paper focuses on the list of 10,000 top‐cited authors in computer science available as part of CiteSeer. Using data from several consecutive lists a model of how authors accrue citations with time is constructed. By comparing the rate at which individual authors accrue citations with the average rate, predictions are made of how their ranking in the list will change in the future.