Probability distributions and inequality measures for analyses of circulation data

Abstract

Bibliometric analysis of circulation data is a useful tool for effective library management. For the purpose of such an analysis, the theoretical basis is provided by the study of the distribution of transactions. The negative binomial distribution is a good approximation to describe the circulation distribution. Thus, the phenomenon of the distribution of the transactions is a manifestation of success-breeds- success phenomenon. However, for highly skewed, long tailed, heterogeneous data, the negative binomial is unlikely to be a good approximation. In such cases, one may derive a distribution function by maximizing the entropy. Further, various measures of inequality are found to be equally useful in analyzing the data. Amongst these, Gini's index, Lorenz ratio, information measures (such as Theil's measure, entropy, etc.) are mostly used and easily interpretable measures. Analysis of circulation data helps us in effective physical organization of documents, in file organization, in acquiring multiple copies of documents and in determing the core collection.