Some comments on “The estimation of lost multi-copy documents: A new type of informetrics theory” by Egghe and Proot

摘要

Egghe and Proot [Egghe, L., & Proot, G. (2007). The estimation of the number of lost multi-copy documents: A new type of informetrics theory. Journal of Informetrics] introduce a simple probabilistic model to estimate the number of lost multi-copy documents based on the numbers of retrieved ones. We show that their model in practice can essentially be described by the well-known Poisson approximation to the binomial. This enables us to adopt a traditional maximum likelihood estimation (MLE) approach which allows the construction of (approximate) confidence intervals for the parameters of interest, thereby resolving an open problem left by the authors. We further show that the general estimation problem is a variant of a well-known unseen species problem. This work should be viewed as supplementing that of Egghe and Proot [Egghe, L., & Proot, G. (2007). The estimation of the number of lost multi-copy documents: A new type of informetrics theory. Journal of Informetrics]. It turns out that their results are broadly in line with those produced by this rather more robust statistical analysis.