Revisiting citation aging: a model for citation distribution and life-cycle prediction

The study of citation distribution provides retrospective and prospective picture of the evolving impact of a corpus of publications on knowledge community. All distribution models agree on the rise of the number of citations in the first years following the publication to reach a peak and then tend to be less cited when time passes. However, questions such as how long it will continue being cited and what is objectively the rate of the decline remain unanswered. Built up of simple polynomial function, the proposed model is proven to be suitable to represent the observed citation distribution over time and to interestingly identify with accuracy when the major loss of citations happens. I calculate from the model the ‘residual citations’ representing the citations kept after a long time period after publication year. I demonstrate that the residual citations may be greater than or equal to zero, meaning that the ‘life-cycle’ of the corpus is infinite, contrary to what some researches termed to be around 21 years. This model fits the observed data from SCI according to R-sq which is greater than 98.9%. Rather, it is very simple and easy to implement and can be used by not highly-skilled scientometric users. Finally, the model serves as a citation predictive tool for a corpus by determining the citations that would obtain at any time of its life-cycle.