Some practical aspects of fitting and testing the Zipf-Mandelbrot model

Summary

The standardization of distribution fitting procedures is recommended also in informetrics. We examined the possibility of that standardization when fitting the Zipf-Mandelbrot (ZM) distribution. After propositions of possible steps of standardization, we stress the unique role of maximum likelihood estimates concerning the chi-square goodness-of-fit tests. We touch upon the possible correlation between the parameters of the ZM distribution. A numerical example demonstrates the method and the results.