Abstract | Citation patterns are important to understanding the spread of technological ideas as science is essentially a cumulative activity. One feature only now being appreciated is the obsolescence or ageing patterns in citations and the insights their study can bring. There have been a number of studies examining, predicting, modeling and plotting citation delays, ageing and the publication cycle. These normally apply lognormal, log-logistic and Weibull distributions to scientific papers. This paper adds to this body of work by examining a set of 18 other distributions, and tests their predictive power on a new data set based on 10 years of ISI Citation data for 10 innovation centered journals. The resulting grouping of journals appears to be a useful proxy for academic-practitioner involvement and warrants further investigation. The finding that the three-parameter Inverse Gaussian provides the best fit to the data extends the understanding of this process. As well as allowing the classification of literature, this improved representation of citation obsolescence will allow us to predict earlier and more precisely those scientific ideas which are generating noteworthy attention or may be suitable for early exploitation. |
---|